WO2020134236A1 - 收割机及其自动驾驶方法 - Google Patents

收割机及其自动驾驶方法 Download PDF

Info

Publication number
WO2020134236A1
WO2020134236A1 PCT/CN2019/107551 CN2019107551W WO2020134236A1 WO 2020134236 A1 WO2020134236 A1 WO 2020134236A1 CN 2019107551 W CN2019107551 W CN 2019107551W WO 2020134236 A1 WO2020134236 A1 WO 2020134236A1
Authority
WO
WIPO (PCT)
Prior art keywords
image
harvester
area
information
acquisition device
Prior art date
Application number
PCT/CN2019/107551
Other languages
English (en)
French (fr)
Inventor
吴迪
王清泉
沈永泉
王波
张虓
童超
范顺
陈睿
Original Assignee
丰疆智能科技股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority claimed from CN201811638418.XA external-priority patent/CN109588107A/zh
Priority claimed from CN201822267500.8U external-priority patent/CN209983105U/zh
Application filed by 丰疆智能科技股份有限公司 filed Critical 丰疆智能科技股份有限公司
Priority to JP2021538493A priority Critical patent/JP2022516898A/ja
Publication of WO2020134236A1 publication Critical patent/WO2020134236A1/zh

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
    • A01D—HARVESTING; MOWING
    • A01D41/00—Combines, i.e. harvesters or mowers combined with threshing devices
    • A01D41/02—Self-propelled combines
    • A—HUMAN NECESSITIES
    • A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
    • A01D—HARVESTING; MOWING
    • A01D41/00—Combines, i.e. harvesters or mowers combined with threshing devices
    • A01D41/12—Details of combines
    • A01D41/127—Control or measuring arrangements specially adapted for combines

Definitions

  • the invention relates to the field of automatic driving of agricultural machinery, in particular to a harvester and its automatic driving method.
  • the harvester is a crop harvesting machine that harvests grains and straw of crops such as rice and wheat.
  • the harvester also includes a lawn mower to harvest other crops, such as machinery and equipment for harvesting vegetables and fruits.
  • the grain harvester is an integrated machine for harvesting crops. It can complete the harvesting, threshing, and concentrate the grains in the storage bin at one time, and then transport the grains to the transport vehicle through the conveyor belt.
  • the fruit and vegetable harvesting equipment can harvest vegetables and fruits in the farmland at one time, separate the harvested fruits from the stalks, and then sort them.
  • Agricultural machinery and equipment need to consider many factors such as the operated area, unoperated area, and the boundary between the world and the earth when working in the farmland, and during the operation process, it is necessary to adjust the operation of the vehicle and adjust the operation in real time according to the conditions of the crops. parameter. Due to the need to consider the complex operating environment during driving, the prior art agricultural equipment also requires the operator to adjust the operation of the agricultural machinery equipment based on real-time farm crop information. The probability of occurrence of a judgment error in the operation of the agricultural machinery equipment controlled by manual operation is large, and the probability of failure of the machinery equipment during the operation is large.
  • This prior art harvester has at least one of the following defects: First, when the harvester is in operation, the vibration of the vehicle itself and the unevenness of the farmland and land will cause the harvester body to shake up and down, resulting in the installation of The camera device of the harvester body cannot capture images at stable positions. Therefore, the images acquired through the camera device are often blurred, and cannot provide information support for intelligent operations and automatic driving. Secondly, the prior art camera device is fixedly installed on the harvester body, and can only acquire images in a single direction, such as the image in front of the harvester, but cannot adjust the shooting direction of the camera device according to the situation And location.
  • the prior art mobile camera equipment or fixed camera equipment such as a drone camera device or a camera device fixed in the farmland, captures the image around the harvester and transmits it to the harvester body for the The harvester body reads the image captured by the camera device.
  • the problem of unclear image capture is solved to some extent, images captured based on the camera device itself or based on the position of the drone cannot be obtained from the perspective of the harvester itself. Therefore, the acquired image cannot be well recognized.
  • the agricultural machinery and equipment in the prior art usually cause errors in the operation due to the inaccurate set operation path, and even serious mechanical failures.
  • the satellite positioning method using PTK has high requirements for the performance of agricultural equipment, and the manufacturing cost and maintenance cost required are relatively high. Therefore, this prior art automatic driving positioning method is not applicable to the current In the automatic driving mode of agricultural machinery and equipment.
  • a main advantage of the present invention is to provide a harvester and its automatic driving method, wherein the harvester recognizes the area of the farmland in the graphic based on the captured at least one visual image.
  • Another advantage of the present invention is to provide a harvester and its automatic driving method, wherein the harvester recognizes crop information such as the type, height, and maturity of the crop in the graphic based on the captured at least one visual image.
  • Another advantage of the present invention is to provide a harvester and its automatic driving method, wherein the harvester recognizes the unoperated area, the operated area, and the field boundary area in the visual image based on the visual image, so that The identified area controls the driving path of the harvester.
  • Another advantage of the present invention is to provide a harvester and its automatic driving method, wherein the harvester recognizes the information of the crop in the image based on the visual image, and the harvester recognizes the information in the image Adjust the operating parameters of the harvester to improve the operating quality and efficiency of the harvester.
  • Another advantage of the present invention is to provide a harvester and its automatic driving method, wherein the image acquisition device of the harvester is a pan-tilt camera device, wherein the pan-tilt camera device has an anti-shake shooting function, which improves the harvesting The accuracy and stability of the machine to obtain visual images.
  • Another advantage of the present invention is to provide a harvester and its automatic driving method, wherein the image acquisition device is provided on a harvester body of the harvester, wherein the harvester photographs the image through the image acquisition device Images around the harvester body.
  • Another advantage of the present invention is to provide a harvester and its automatic driving method, wherein the image acquisition device is provided on a harvester body of the harvester, wherein the image acquisition device is provided on the harvester body , wherein the image acquisition device shoots at least one visual image or visual image based on the position of the field of view of the harvester body, so as to identify the information around the harvester body according to the captured image information.
  • Another advantage of the present invention is to provide a harvester and its automatic driving method, wherein the image acquisition device can be adjusted to take images of different angles and different directions based on the position of the main machine of the harvester to facilitate acquisition of Describe the images of the harvester main machine in different directions.
  • Another advantage of the present invention is to provide a harvester and its automatic driving method, wherein the image acquisition device is a mechanical gimbal camera or an electronic gimbal camera, and the stability of the visual image is improved by the image acquisition device Sex.
  • Another advantage of the present invention is to provide a harvester and its automatic driving method, wherein a path planning system of the harvester automatically plans a route based on the current vehicle positioning information, the information recognized by the image processing system, and the information of the navigation system.
  • Another advantage of the present invention is to provide a harvester and its automatic driving method, wherein the harvester plans the driving path and working route of the harvester based on the area identified by the visual image.
  • Another advantage of the present invention is to provide a harvester and its automatic driving method, wherein an image acquisition device of the harvester acquires the visual image of the surrounding farmland in real time, and updates the path navigation information planned by the harvester in real time .
  • Another advantage of the present invention is to provide a harvester and its automatic driving method, wherein the harvester captures an image in real time through the image acquisition device, recognizes the area in the visual image, and changes according to the area Update or adjust the working route of the harvester in real time to improve the working quality of the harvester.
  • Another advantage of the present invention is to provide a harvester and its automatic driving method, wherein the image processing system of the harvester uses image segmentation technology to identify the unworked in the image based on the acquired visual image information Area, the operated area, and the field boundary area, and the boundary dividing the two adjacent areas.
  • Another advantage of the present invention is to provide a harvester and its automatic driving method, wherein the image processing system of the harvester uses image segmentation technology to identify the type and height of crops in the image based on the acquired visual image information Plant information, such as grain fullness, for the operating system of the harvester to adjust the operating parameters based on the crop information.
  • Another advantage of the present invention is to provide a harvester and its automatic driving method, wherein the image processing system of the harvester recognizes the area boundary in the image based on the acquired image information, so that the path planning system is based on The identified boundary of the area plans a driving path of the vehicle.
  • Another advantage of the present invention is to provide a harvester and its automatic driving method, wherein the harvester does not require high-precision satellite positioning, which reduces the difficulty of manufacturing the automatic driving equipment and reduces the maintenance cost of the equipment .
  • Another advantage of the present invention is to provide a harvester and its automatic driving method, wherein the harvester performs path planning based on the area division information output by the image processing system to realize automatic driving and automatic driving operations.
  • a harvester of the present invention that can achieve the foregoing objectives and other objectives and advantages includes:
  • At least one image acquisition device wherein the image acquisition device is provided on the harvester host, and the image acquisition device captures images around the harvester host, and
  • An image processing system wherein the image processing system recognizes farmland information in the image based on an image captured by the image acquisition device, wherein the harvester host automatically determines the farmland information identified by the image processing system Control driving.
  • the harvester further includes a path planning system, wherein the path planning system plans at least one travel planning path based on the farmland information identified by the image processing system, wherein the harvester The host computer controls driving according to the travel planning path planned by the path planning system.
  • the image processing system uses image segmentation and recognition technology to identify the information of the farmland in the image, and plans the area of the farmland in the image based on the identified information.
  • the image processing system uses image segmentation and recognition technology to identify crop information in the image for the harvester host to automatically adjust operation parameters based on the identified information.
  • the image acquisition device is an anti-shake PTZ camera device
  • the image acquisition device is loaded on the harvester host
  • the photo is taken in a photographic manner based on the position of the harvester host Images around the harvester mainframe.
  • the image acquisition device is a mechanical anti-shake gimbal device
  • the image acquisition device includes a gimbal and at least one camera, wherein the gimbal mounts the camera to the harvester
  • the host the camera is installed on the gimbal, and the camera is supported by the gimbal to maintain balance.
  • the image acquisition device is an electronic pan/tilt device, and the image acquisition device controls the angle of view and zoom of the lens, thereby preventing the camera of the image acquisition device from shaking.
  • the image acquisition device is disposed at the front of the harvester host, at the top of the harvester host, on the left or right side of the harvester host, or the harvester The rear of the host.
  • the image processing system further includes:
  • An image segmentation module wherein the image segmentation module divides the image into a plurality of pixel regions, wherein each of the pixel regions includes at least one pixel unit;
  • a characterization module wherein the characterization module extracts features corresponding to each pixel region based on the pixel unit of the pixel region;
  • An area dividing module wherein the area planning module identifies and divides the area of the image according to the characteristics of the pixel area.
  • the harvester further includes a positioning device and a navigation system, the positioning device and the navigation system are provided on the harvester host, wherein the positioning device acquires the harvester Position information of the host, wherein the navigation system provides navigation information for the grain processing body.
  • the path planning system further includes:
  • An operation area setting module wherein the operation area setting module sets the operation area of the farmland and the operation boundary obtained from the boundary area of the farmland;
  • a driving path planning module wherein the positioning information based on the harvester host, the image processing system recognizes the area planning information of the image, and the navigation information of the navigation system to obtain at least one driving planning path.
  • the harvester host includes a vehicle body, at least one operating system provided on the vehicle body, and a driving control system, the vehicle body drives the operation system to operate, wherein the The driving control system controls the operation of the vehicle body and the operating parameters of the operating system.
  • the driving control system acquires the information of the image captured by the image acquisition device recognized by the image processing system, automatically controls the driving route of the vehicle body and controls the operation of the operating system Parameters to achieve unmanned automatic driving and harvesting operations.
  • the present invention further provides an automatic driving method of a harvester, wherein the automatic driving method includes the following steps:
  • step (a) of the above-mentioned automatic driving method further includes: identifying information of corresponding crops in the farmland in the image, wherein the information of the crops includes types of crops, height of crops, and fullness of particles And other information.
  • step (b) of the above-mentioned automatic driving method further includes steps:
  • (b.2) Plan at least one travel planning route based on the identified area.
  • the step (b.1) of the above-mentioned automatic driving method further includes the steps of segmenting the image by using an image segmentation technique, and identifying a region that divides the image.
  • the image processing system uses image segmentation technology to segment the image information, and recognizes that the area dividing the image is the unworked area, The operated area and the field boundary area.
  • step (b.1) of the automatic driving method further includes the following steps:
  • the classification label of the image is output.
  • the step (b.2) of the above-mentioned automatic driving method further includes the step of: based on the positioning information of the harvester host, the area planning information of the image, and the navigation information of the navigation system, Plan out the driving plan path.
  • the above-mentioned automatic driving method further includes: step (b.3) comparing whether the area division and the area boundary range identified by the image processing system are consistent with the previous area boundary range, if not , The area division and the area boundary range corresponding to the image are adjusted, and if the consistency can be maintained, the area division and the boundary range are kept unchanged.
  • the above-mentioned automatic driving method further includes the step of: (d) adjusting the operating parameters of the operating system of the harvester host based on the identification information of the image.
  • FIG. 1 is a system schematic diagram of a harvester according to the first preferred embodiment of the present invention.
  • FIG. 2 is a schematic diagram of image acquisition of the harvester according to the above-mentioned preferred embodiment of the present invention.
  • FIG. 3A is a schematic diagram of an image acquired by the harvester according to the above preferred embodiment of the present invention.
  • 3B is a schematic diagram of another image acquired by the harvester according to the above-described preferred embodiment of the present invention.
  • 3C is a schematic diagram of another image acquired by the harvester according to the above-described preferred embodiment of the present invention.
  • FIG. 4 is a schematic diagram of an image processing system of the harvester according to the above preferred embodiment of the present invention for dividing and identifying the image area.
  • 5A is a schematic diagram of the image processing system of the harvester according to the above preferred embodiment of the present invention dividing the image area.
  • 5B is a system block diagram of the image processing system of the harvester according to the above-described preferred embodiment of the present invention.
  • FIG. 6 is a schematic diagram of the image processing system of the harvester according to the above preferred embodiment of the present invention extracting the image region feature recognition.
  • FIG. 7 is a schematic diagram of the area division of the image processing system of the harvester according to the above-mentioned preferred embodiment of the present invention outputting the image.
  • FIG. 8 is a schematic diagram of the change of the boundary division of the area division of the image output system of the harvester according to the above preferred embodiment of the present invention.
  • FIG. 9 is a schematic diagram of an automatic driving scene of the harvester according to the above preferred embodiment of the present invention.
  • FIG. 10 is a system schematic diagram of a path planning system of the harvester according to the above preferred embodiment of the present invention.
  • 11A is a schematic diagram of farmland path planning generated by the path planning system of the harvester according to the above-described preferred embodiment of the present invention.
  • FIG. 11B is a schematic diagram of the driving path adjusted by the path planning system of the harvester according to the above preferred embodiment of the present invention.
  • FIG. 12 is a schematic diagram of the overall structure of a harvester with an image acquisition device according to a second preferred embodiment of the present invention.
  • FIG. 13 is a schematic diagram of an image captured by the image acquisition device of the harvester according to the above preferred embodiment of the present invention.
  • FIG. 14 is a schematic structural view of the image acquisition device of the harvester according to the above-described preferred embodiment of the present invention, wherein the image acquisition device is implemented as a mechanical pan-tilt device.
  • 15 is a schematic diagram of the installation position of the image acquisition device of the harvester according to the above-described preferred embodiment of the present invention.
  • 16A is a schematic diagram of an image processing system of the harvester according to the above-mentioned preferred embodiment of the present invention identifying a farmland area in an image captured by the image acquisition device.
  • 16B is a schematic diagram of an image processing system of the harvester according to the above preferred embodiment of the present invention identifying crops in an image captured by the image acquisition device.
  • FIG. 17 is a schematic diagram of another optional implementation manner of the image acquisition device of the harvester according to the above-described preferred embodiment of the present invention, wherein the image acquisition device is implemented as an electronic pan/tilt device.
  • the term “a” should be understood as “at least one” or “one or more”, that is, in one embodiment, the number of an element can be one, and in other embodiments, the The number can be more than one, and the term “one” cannot be understood as a limitation on the number.
  • FIGS. 1 to 9 of the accompanying drawings of the description of the present invention a harvester and its automatic driving method according to the first preferred embodiment of the present invention are disclosed and explained in the following description, wherein the harvester can be It is implemented as crop harvester equipment with grain processing functions, vegetable, fruit harvesting equipment, lawn mowing equipment, and other types of harvesting devices. It can be understood that the type of harvester described in the present invention is only of an exemplary nature, not a limitation.
  • the harvester acquires at least one image in the surroundings, and processes the area type of the farmland in the image based on visual recognition, and divides various area types and boundaries of the farmland in the image.
  • the harvester divides the type and boundary of each area according to the division, wherein the area type of the farmland divided by the harvester includes at least one operated area 100, at least one unoperated area 200, and at least one field boundary area 300, and the harvest According to the type of the divided area, the aircraft plans the walking route of the vehicle by the navigation system to realize unmanned automatic driving and unmanned automatic driving.
  • the image acquired by the harvester of the present invention is image data information corresponding to crop grains in farmland, wherein the image is an image of the periphery of the vehicle acquired based on the current position of the vehicle.
  • the harvester does not need satellite positioning information with too high accuracy, and only needs satellite positioning with ordinary meter accuracy (GPS positioning or Beidou positioning, etc.).
  • the image acquired and processed by the harvester is different from the self-driving car, therefore, the path planning and driving manner formed by the harvester are also different.
  • the harvester of the present invention recognizes the area of the farmland based on vision and the automatic driving function is different from the recognition mode of the automatic driving car.
  • the harvester acquires at least one image around the area, wherein the harvester divides the area type corresponding to the farmland and the boundary between the areas according to the acquired image recognition.
  • the harvester obtains the images around the harvester by fixed-point photographing, video shooting, mobile photographing, and the like. It can be understood that the manner in which the harvester acquires the image is only exemplary in nature, and not limiting.
  • the harvester includes a harvester host 10 and at least one image acquisition device 20, wherein the image acquisition device 20 acquires at least one image around the harvester host 10.
  • the image acquisition device 20 is provided on the harvester host 10, wherein the image acquisition device 20 acquires the images around the harvester host 10 by way of photographing or video shooting. More preferably, the image acquisition device 20 is provided in front of the harvester host 10, wherein the image acquisition device 20 can acquire the image in front of the harvester host 10 in real time, wherein the harvester host 10
  • the travel route is set based on the area identified by the image information captured by the image acquisition device 20. It is worth mentioning that the image captured by the image acquisition device 20 is based on the image within the field of view of the harvester host 10. In other words, the image acquisition device 20 acquires an image based on the direction of the field of view of the harvester host 10, and adjusts the traveling direction of the harvester host 10 according to the position where the image acquisition device 20 is mounted to the harvester host 10.
  • the image acquisition device 20 captures the vision of the world in the driving direction of the harvester host 10, wherein the image may be a two-dimensional planar image or a three-dimensional stereoscopic image that is captured. It can be understood that the type of the image captured by the image acquisition device 20 is merely an example, not a limitation.
  • the harvester host 10 is implemented as a grain harvester device, wherein the harvester host 10 is controlled to travel to an unworked area of farmland 200 performs a harvesting operation to harvest crops in the non-operational area 200, such as rice, wheat, corn, etc.
  • the harvester host 10 performs automatic driving in the field according to the area divided by the image obtained by the image obtaining device 20, and unmanned automatic driving. It can be understood that the type of the harvester main machine 10 is merely an example, not a limitation.
  • FIG. 3A shows the image captured by the image acquisition device 20 when the harvester host 10 is used as a grain harvester.
  • the area in the farmland is divided into at least one unharvested area 100a, at least one harvested area 200a, and at least one field boundary area 300a according to whether the cereal is harvested, wherein the harvested area 200a is an area where crops have been harvested, wherein The original crops in the harvested area 200a are harvested.
  • the unharvested area 100a is an area where crops still exist, and there are still growing crops in the unharvested area 100a.
  • the field boundary area 300a is a ridge in the farmland that separates the interval between crops, an outer boundary around the farmland, and an area where obstacles exist in the farmland, wherein the field boundary area 300a is not planted with crops.
  • FIG. 3B shows the image captured by the image acquisition device 20 when the harvester host 10 is used as a mowing device.
  • the area in the farmland is divided into at least one unharvested area 100b, at least one harvested area 200b, and at least one field boundary area 300b according to whether the grain is cultivated, wherein the unharvested area 100b represents an area that has not been harvested crops, so
  • the harvested area 200b represents an area where crops have been cultivated, and the field boundary 300b is an outer boundary that separates the ridges where crops are planted and the periphery of the farmland, and the area where obstacles exist in the farmland.
  • FIG. 3C shows the image captured by the image acquisition device 20 when the harvester host 10 is used as a stalk plant or fruit harvesting device, such as a vegetable harvester device.
  • the area in the farmland is divided into at least one unharvested area 100c, at least one harvested area 200c, and at least one field boundary area 300c according to whether the grain is sprayed.
  • the unharvested area 100c represents an area where crops have not yet been harvested
  • the harvested area 200c represents an area of crops that have been harvested
  • the field boundary 300b is an outer boundary that separates the ridges where crops are planted and the periphery of the farmland, and There are obstacles in the farmland.
  • the images acquired by the image acquisition device 20 are identified by the image segmentation recognition technology to the unoperated area 100, the operated area 200, and the field boundary area 300, and distinguished The boundary between the areas.
  • the harvester further includes an image processing system 30, wherein the image processing system 30 recognizes the unworked from the image using image segmentation recognition technology based on the image of the farmland acquired by the image acquisition device 20 The area 100, the operated area 200, and the field boundary area 300.
  • the image processing system 30 uses image segmentation recognition technology to identify the areas and boundaries in the image that are used to represent the areas and boundaries of the farmland in front of the harvester main machine 10 traveling. Based on the regions and boundaries identified by the image processing system 30 using image segmentation recognition technology, the harvester main unit 10 is controlled to travel and perform operations in an unworked area of farmland.
  • a harvester device the image acquisition device 20 provided at the front end of the harvester device acquires an image of farmland in front of the harvester device, wherein the image captured by the image acquisition device 20 is divided and recognized by the image processing system 30 To identify the unworked area 100, the worked area 200, and the field boundary area 300 of the farmland that divides the farmland in the traveling direction of the harvester device.
  • the harvester host 10 which is the host of the harvester device, plans the vehicle travel path and harvesting operation based on the area and boundary recognized by the image processing system 30.
  • the image processing system 30 uses image segmentation and recognition technology to identify the types of crops in the image provided by the image acquisition device 20, the height of the crops, and the fullness of crop particles.
  • the image processing system 30 can determine whether the crop has been harvested based on the identified type of crop in the image and the height of the crop, and can be used to adjust the job based on the fullness information of the identified crop particles in the image parameter.
  • the image processing system 30 can identify the area type and boundary of the farmland according to the image provided by the image acquisition device 20, and can also identify the type, height, grain fullness, and crop maturity of the crops in the farmland.
  • the image processing system 30 is selected from any of the segmentation recognition methods selected from threshold-based segmentation methods, area-based segmentation methods, edge-based segmentation methods, and specific theory-based segmentation methods.
  • the image acquired by the acquiring device 20 performs segmentation recognition to identify regions and boundaries in the image.
  • the image processing system 30 utilizes a deep learning algorithm to recognize the image segmentation and perform area division and boundary definition on the image.
  • the image processing system 30 uses a deep learning algorithm to identify the area and boundary of the corresponding farmland in the image for the harvester host to travel and perform operations based on the identified area and boundary.
  • the deep learning algorithm used by the image processing system 30 is a convolutional neural network algorithm image segmentation recognition technology to identify the unoperated area 100, the operated area 200, and the Described field boundary area 300.
  • the processing algorithm utilized by the image processing system 30 is merely exemplary, not limiting. Therefore, the image processing system 30 can also use other algorithms to segment and identify the acquired image to identify the area and boundary of the farmland in the image.
  • the image processing system 30 divides the image acquired by the image acquisition device 20 into a plurality of pixel regions 301, each of which includes at least one pixel unit. It can be understood that the image corresponds to an area around the harvester main body 10, and accordingly, the pixel area 301 of the image corresponds to image information of a specific area of farmland or crops in the farmland being photographed .
  • Each pixel region 301 formed by division is subjected to a normalization process, so that the pixel unit of the pixel region 301 is normalized into a numerical value or an array having a size corresponding to the pixel value.
  • the image processing system 30 normalizes the divided pixel regions 301 into corresponding numerical values or arrays for the image processing system to extract the features of the image and divide the regions.
  • the image processing system 30 extracts image features corresponding to the pixel region 301 based on the array corresponding to each pixel region 301.
  • the image processing system 30 obtains image features corresponding to the pixel region 301 according to the array corresponding to the pixel region 301.
  • the image processing system 30 uses a convolutional neural network algorithm, such as a two-dimensional convolutional neural network
  • the input layer of the convolutional neural network corresponds to the corresponding two-dimensional array or three-dimensional array in the pixel region 301.
  • the hidden layer of the convolutional neural network performs feature extraction on the array of the input layer, and performs feature selection and information filtering after feature extraction.
  • the convolutional neural network outputs a classification label of the pixel region 301 based on the features corresponding to the array, wherein the classification labels respectively correspond to the unoperated region 100, the operated region 200, and the field boundary region 300.
  • the image processing system 30 recognizes the region features corresponding to the pixel region 301 by extracting the features of the array corresponding to the pixel region 301, wherein the features corresponding to the pixel region 301 It mainly includes the height characteristics of crop plants, the interval of crop plants in farmland, the color of crops, the color of farmland land, the characteristics of crop types, the characteristics of farmland land, the fullness of crop particles, and the number of crop particles.
  • the image processing system 30 outputs a classification label corresponding to the pixel area 301 according to the extracted features, wherein the classification label correspondingly identifies the area type and the boundary line corresponding to the pixel area 301 based on the feature information.
  • the image processing system 30 includes an image segmentation module 31, a characterization module 32, and an area division module 33.
  • the image segmentation module 31 acquires an image captured by the image acquisition module 20, and performs image segmentation processing to form a plurality of the pixel regions 301, where each of the pixel regions 301 includes at least one pixel unit.
  • the feature module 32 uses a deep learning algorithm to extract the feature type corresponding to the pixel region 301, and select features and filter information.
  • the area dividing module 33 divides the image based on the features corresponding to the pixel area 301 extracted by the characterization module 32 to generate the corresponding unoperated area 100, the operated area 200, and the field The classification label of the boundary area 300.
  • the image segmentation module 31 divides the image into a plurality of pixel regions 301, wherein each pixel region 301 has the same size, shape and range. It can be understood that the image segmentation module 31 may also perform segmentation according to the image pixel threshold size. In other words, the size, shape, and range of the pixel region 301 segmented by the degraded segmentation module 31 may be different. More preferably, when the characterization module 32 of the image processing system 30 adopts a convolutional neural network algorithm, the pixel region 301 divided by the image division module 31 is a single pixel unit.
  • the characterization module 32 includes a pixel processing module 321, a feature extraction module 322, and a feature output module 323, wherein the pixel processing module 321 processes the array corresponding to the pixel units in the pixel region 301.
  • the pixel processing module 321 normalizes the pixel area 301 into an array suitable for processing.
  • the feature extraction module 322 inputs the array of the pixel region 301 processed by the pixel processing module 321, extracts the feature type corresponding to the array, and selects the features, and filters the information to retain the available data and discharge Disturb the data, thus making the feature extraction results more prepared.
  • the feature output module 323 outputs the features extracted by the feature extraction module 322 in combination with the feature extraction module 322, and combines the features output by the feature output module 323 by the area division module 33 to generate the Classification label.
  • the area dividing module 33 divides each area corresponding to the image and sets an area boundary based on the features corresponding to the pixel area 301 extracted by the characterization module 32.
  • the area dividing module 33 further includes an area dividing module 331 and a boundary dividing module 332, wherein the area dividing module 331 divides different areas according to the characteristics of the pixel area 301, wherein the boundary dividing module 332 Divide the boundary range corresponding to the area, so as to identify the range of the area.
  • the image acquisition device 20 acquires images in the field of view in front of the harvester host 10 in real time. Accordingly, the image processing system 30 acquires the image captured by the image acquisition device 20 in real time, and uses image segmentation and recognition technology to identify the area division and area boundary range of the image corresponding to the farmland. When the area division and area boundary range identified by the image processing system 30 cannot be consistent with the previous area boundary range, the area division and area boundary range corresponding to the image are adjusted.
  • the image processing system 30 updates the area division and area boundary range corresponding to the image in real time.
  • the harvester further includes a positioning device 40 and a navigation system 50, wherein the positioning device 40 is disposed on the harvester host 10 to obtain the position of the harvester host 10 information.
  • the positioning device 40 uses satellite positioning information to obtain the position information of the harvester host 10, such as a GPS or a Beidou positioning device.
  • the navigation system 50 is provided in the harvester host 10, wherein the navigation system 50 navigates the harvester host 10 for the positioning of the harvester host 10 based on the positioning device 40
  • the information, the area planning information obtained by the image processing system 30, and the navigation information of the navigation system 50 realize unmanned automatic driving and operation.
  • the image of the farmland area division and area boundary range obtained by the image processing system 30 based on the image is updated to the navigation system 50 in real time to update the navigation information of the navigation system 50.
  • the navigation system 50 is implemented as an inertial integrated navigation system. It can be understood that the type of the navigation system 50 is merely an example, not a limitation, and therefore, the navigation system 50 may also be implemented as other types of navigation devices.
  • the harvester main body 10 of the harvester includes a vehicle body 11, an operating system 12 provided on the vehicle body 11, and a driving control system 13, wherein the operating system 12 is controlled by the vehicle
  • the main body 11 drives and implements grain processing operations, such as harvesting operations.
  • the driving control system 13 controls the running of the vehicle body 11 and controls the operation of the working system 12. It is worth mentioning that the driving control system 13 has an unmanned driving mode and an operating driving mode. When the harvester is in the unmanned driving mode, the driving control system 13 controls the vehicle body 11 to automatically operate and the operation of the operation system 12. Accordingly, when the harvester is in the operation driving mode, the driving control system allows the driver to manually operate the vehicle body 11 and control the operation of the operation system by manual operation.
  • the harvester is a harvester device
  • the operating system 12 is implemented as a harvesting device.
  • the driving control system 13 controls the running of the vehicle body 11 and controls the operation of the working system 12. In other words, the driving control system 13 controls the adjustment of the operating parameters of the operating system 12 while the vehicle body 11 is traveling.
  • the driving control system 13 acquires the information of the image processing system 30 to identify the types of crops, the height of the crops, the degree of grain fullness, the diameter of the crop stalks, etc. in the image, and adjusts the operation based on the acquired information
  • the operating parameters of the system 12 are, for example, adjusting the operating speed of the operating system 12, the width of the operation, the height of the operation, and adjusting the parameters of the off-force processing.
  • Figure 9 of the accompanying drawings of the present specification shows an embodiment of the unmanned driving and harvesting operation of the harvester in farmland.
  • the driving control system 13 of the harvester host 10 acquires the positioning information of the vehicle body 11 provided by the positioning device 40, the navigation system 50 The provided navigation information and the area identification information provided by the image processing system 30 further control the vehicle body 11 to travel in the unoperated area 100 of the farmland to complete the grain harvesting operation.
  • the image acquisition device 20 acquires the image of the vehicle body 11 in front of the vehicle in real time, wherein the image is recognized by the image processing system 30 using image segmentation recognition technology Out of the area and boundary.
  • the image processing system 30 replaces the original area division and boundary range, and updates the navigation of the navigation system 50 Data to enable the driving control system 13 to acquire new navigation information to adjust the driving and working route.
  • the harvester is based on the position information of the harvester host 10 acquired by the positioning device 40, the area planning information of the image recognized by the image processing system 30, and the The navigation information of the navigation system 50 generates at least one planned route.
  • the driving control system 13 of the harvester host 10 controls the driving of the vehicle body 11 and the operation of the operation system 12 according to the generated planned route.
  • the harvester further includes a path planning system 60, wherein the path planning system plans at least one vehicle's travel path for the harvester host 10.
  • the path planning system 60 obtains the positioning information of the positioning device 40, obtains the area planning information of the image recognized by the image processing system 30, and obtains the navigation information of the navigation system 50, and plans based on the obtained information The travel path of the vehicle body 11.
  • the path planning system 60 identifies or sets at least one operation area 601 and operation boundary 602 corresponding to the farmland, where the operation area 601 is the largest operation of the harvester Range, wherein the driving control system 13 controls the vehicle body 11 to travel within the range of the working boundary 602.
  • the working area 601 and the working boundary 602 can be identified by the image processing system 30 by identifying the field boundary area 300 in the image by identifying the maximum area range and boundary of the working area 601.
  • the path planning system 60 sets the working area 601 of the harvester in a setting manner.
  • the path planning system 60 plans at least one travel path based on the outermost working boundary 602 of the working area 601. When the width of the work area 601 is greater than the work width of the work system 12, the path planning system 60 plans a "back"-shaped travel route, or an "S"-shaped travel route. It can be understood that the manner in which the driving route planned by the path planning system 60 is merely exemplary, and not limiting. Therefore, other driving routes can also be applied here.
  • the path planning system 60 replans at least one travel path based on the range of the current non-work area 100.
  • the path planning system 60 updates the work area 601 and the work boundary 602 for the vehicle body 11, and according to the update A new driving path is planned for the working area 601 of.
  • the driving control system 13 controls the vehicle body 11 to travel according to the travel path planned by the path planning system 60, wherein the driving control system 13 controls the working system 12 to harvest the working area 401 The outermost crop. In other words, the driving control system 13 controls the working system 12 to harvest crops in the unworked area 100 based on the working boundary 602.
  • the path planning system 60 of the harvester includes a work area setting module 61, a driving path planning module 62, and a path adjustment module 63.
  • the working area setting module 61 recognizes the working area 601 and the working boundary 602 of the farmland based on the image processing system 30 identifying the boundary area of the farmland in the image; or setting the harvesting by setting
  • the main machine 10 operates the operation area 601 and the operation boundary 602 in the farmland. Since the operation of the harvester host 10 causes the unoperated area 100 and the operated area 200 to change, the operation area setting module 61 updates the range of the operation area 601 and the boundary of the operation boundary 602 in real time in order to A new said unworked area 100 and said already worked area 200 are generated.
  • the driving path planning module 62 obtains at least one driving planning path based on the positioning information of the harvester host 10, the image processing system 30 identifying the area planning information of the image, and the navigation information of the navigation system 50 603, wherein the driving control system 13 controls the vehicle body 11 to travel according to the travel planning path 603.
  • the path adjustment module 63 adjusts the driving direction of the harvester body 10 based on the information of the image processing system 30 identifying the crops of the image to form a vehicle driving path 604, wherein the vehicle driving paths 604 substantially coincide or Parallel to the driving planning path 603.
  • the vehicle travel path generated by the path adjustment module 63 deviates from the travel planning path 603.
  • the harvester includes a harvester host 10 and at least one image acquisition device 20, wherein the image acquisition device 20 is disposed on the harvester host 10, and the image acquisition device 20 photographs the farmland where the harvester host 10 is located Images or video images for the harvester host 10 to control the driving direction and/or operating parameters based on the image or image information captured by the image acquisition device 20.
  • the image acquisition device 20 captures information on the farmland around the farmland where the harvester host 10 is located based on the position of the harvester host 10.
  • the image acquisition device 20 captures an image in the field of view, such as an image in the field of vision of the driver, so as to adjust the operating parameters of the harvester host 10 according to the captured image, such as adjusting the driving route, Travel speed, operating parameters, etc.
  • the image acquisition device 20 is carried to the harvester host 10, wherein the image and image information captured by the image acquisition device 20 is transmitted to the harvester host 10 for the The harvester host 10 adjusts the operating parameters based on the information.
  • the image acquisition device 20 is mounted on the harvester main body 10, wherein the image acquisition device 20 shoots a clear image when the harvester main body 10 shakes.
  • the image acquisition device 20 is an anti-shake camera device, which can avoid mechanical vibration of the harvester main body 10 itself and shaking caused by unevenness of the world during shooting.
  • the harvester host 10 controls the travel path and operation parameters under the operation of an operator or automatically to realize the operation of the harvester. In other words, the harvester host 10 adjusts the operation and operation parameters based on the image information captured by the image acquisition device 20 to achieve precise operation and/or unmanned autonomous driving operation.
  • the image acquisition device 20 is implemented as a pan-tilt camera device, wherein the image acquisition device 20 takes stable quality images in the case of vibration or jitter Or image.
  • the image acquisition device 20 is a mechanical pan/tilt device, wherein the image acquisition device 20 is mounted on the harvester main body 10 by mechanical connection, and the image The acquiring device 20 realizes the anti-shake captured image through a mechanical anti-shake method.
  • the type of the image acquisition device 20 is only exemplary, not limiting. Therefore, other types of structures and installation methods can also be applied here.
  • the image acquisition device 20 includes a pan-tilt head 21 and at least one camera 22, wherein the pan-tilt head 21 installs the camera 22 to the harvester host 10, and the pan-tilt head 21
  • the installation position of the camera 22 is fixed.
  • the bottom end of the pan-tilt head 21 is loaded onto the harvester main body 10, and the pan-tilt head 21 is fixed by the harvester main body 10, wherein the upper end of the pan-tilt head 21 is set to be connected to the camera 22.
  • the camera 22 is supported by the gimbal 21 to maintain relative balance, so as to stably capture images or videos.
  • the camera 22 shoots images or videos around the harvester host 10 under the support of the pan/tilt head 21, wherein the camera 22 shoots the harvester host 10 based on the installation position of the pan/tilt head 21 The image within the field of view.
  • the camera 22 of the image acquisition device 20 acquires at least one visual image by taking pictures based on the position of the harvester host 10.
  • the camera 22 of the image acquisition device 20 acquires the image based on the field of view of the harvester host 10, thereby avoiding image data caused by changes in the position of the camera device 20 and the harvester host 10 Inaccurate question.
  • the image acquired by the harvester of the present invention is image data information corresponding to crop grains in farmland, wherein the image is an image of the periphery of the vehicle acquired based on the current position of the vehicle.
  • the harvester does not need satellite positioning information with too high accuracy, and only needs satellite positioning with ordinary meter accuracy (GPS positioning or Beidou positioning, etc.).
  • the image acquired and processed by the harvester is different from the self-driving car, therefore, the path planning and driving manner formed by the harvester are also different.
  • the harvester of the present invention recognizes the area of the farmland based on vision and the automatic driving function is different from the recognition mode of the automatic driving car.
  • the gimbal 21 of the image acquisition device 20 further includes a gimbal fixing piece 211 and at least one gimbal moving piece 212, wherein the gimbal moving piece 212 is movably connected to the gimbal fixing piece 211 .
  • the pan-tilt fixing piece 211 is fixedly installed on the harvester main body 10, wherein the camera 22 is mounted to the pan-tilt moving piece 212.
  • the pan-tilt moving part 212 of the pan-tilt 21 movably supports the camera 22, so that the camera 22 maintains a stable relative position when the harvester main body 10 shakes, so that a clear image is set.
  • the pan-tilt fixing member 21 of the pan-tilt 21 and the harvester host 10 Simultaneous ground shaking, wherein the pan-tilt moving member 212 of the pan-tilt 21 moves relative to the pan-tilt fixing member 211, neutralizing the vibration generated by the pan-tilt fixing member 211, thereby maintaining the position of the camera 22 Stability.
  • the pan/tilt moving member 212 shakes or vibrates in the up-down direction, left-right direction, and front-to-back direction of the motion fixing member 211 to keep the camera 22 at a stable photographing position, thereby capturing a stable image information.
  • the camera 22 of the image acquisition device 20 is provided to the pan-tilt moving part 212 of the pan-tilt 21, wherein the camera 22 is fixedly or movably mounted to the The pan/tilt moving part 212 of the pan/tilt 21.
  • the camera 22 is movably disposed on the moving member 212, wherein the camera 22 can rotate based on the upper end of the pan/tilt moving member 212 to capture images in different directions of the field of view.
  • the camera 22 is fixedly mounted to the upper end of the pan-tilt moving member 212, wherein the camera 22 takes an image within a specified field of view under the fixed support of the pan-tilt 21, such as a shooting place
  • a specified field of view under the fixed support of the pan-tilt 21, such as a shooting place
  • the image in the field of view in front of the main body 10 of the harvester will be described.
  • the camera 22 includes a camera body 221 and at least one camera driving device 222, wherein the camera driving device 222 drives the movement of the camera body 221 to capture images in different directions.
  • the camera body 221 is movably disposed on the pan/tilt moving member 212, wherein the camera body 221 can be rotated in the up-down direction under the driving action of the camera driving device 222 to photograph the harvester host 10 Images of farmland and crops at distant and nearby locations. It can be understood that when the camera body 221 is driven to rotate downward by the camera driving device 222, the camera body 221 captures an image of the harvester main body 10 in order to clearly identify the crop information in the image . When the camera body 221 is driven by the camera driving device 222 to rotate upward, the camera body 221 captures an image of the harvester main unit 10 at a distance, so as to identify the working area and field of the farmland through the image Border area.
  • the camera driving device 222 drives the camera body 221 to rotate in the left-right direction, so that the camera body 221 can capture left and right images of the harvester main body 10 in order to identify the unworked area 100 of the farmland and Worked area 200, and field border area 300.
  • FIG. 15 of the accompanying drawings of the specification of the present invention several optional installation methods and installation positions of the image acquisition device 20 installed on the harvester host 10 are shown.
  • the image acquisition device 20 of the harvester is disposed at the front side position, upper top end, left side, right side, and rear of the harvester main body 10 Etc. It can be understood that, the installation position of the image acquisition device 20 is different, the captured image is different, and the information recognized from the image is also different.
  • the image acquisition device 20 provided on the front side of the harvester host 10 takes an image in front of the harvester host 10, and when the harvester travels forward, the harvester host The image acquisition device 20 on the front side of 10 captures the working condition of the harvester main machine 10 in order to adjust the traveling path, working parameters, etc. of the harvester main machine 10 according to the photographed working condition.
  • the image acquisition device 20 provided on the rear side of the harvester host 10 captures an image behind the harvester host 10, and when the harvester travels forward, the image capture device 20 captures the image An image of the work area 200. Identify whether the harvesting operation of the harvester host 10 is qualified by identifying the image of the operated area 200 taken by the image acquisition device 20 on the rear side of the harvester host 10, so as to adjust the harvester host 10 job parameters. It can be understood that, through the image captured by the image acquisition device 20 provided on the rear side of the harvester host 10, the harvester host 10 recognizes whether the crops in the working area 200 are completely harvested, and whether crop particles are left behind, etc. . The harvester host 10 is adjusted the operation parameters according to the information identified in the image, thereby improving the harvesting operation. It is worth mentioning that, when driving in reverse, the image captured by the image acquisition device 20 provides the driver with a reverse image.
  • the image acquisition device 20 provided at the top end of the harvester main body 10 takes a long-distance image of the harvester main body 10 so as to recognize the work area of the farmland, the field boundary area, etc. based on the image.
  • the image acquisition device 20 provided at the top end of the main unit 10 of the harvester is a rotatable pan/tilt camera.
  • the image acquisition device 20 provided on the left or right side of the harvester main machine 10 takes an image on the left or right side of the harvester main machine 10. Based on the image on the left or right side of the harvester host 10, the crops in the farmland in the image are identified so as to identify the unoperated area 100, the operated area 200, and the field boundary area 300.
  • the harvester further includes an image processing system 30, a positioning device 40, and a navigation system 50, wherein the image processing system 30, the positioning device 40, and the navigation The system 50 is installed in the harvester main body 10.
  • the positioning device 40 acquires the position information of the harvester host 10 and transmits the acquired position information to the harvester host 10.
  • the navigation system 50 provides navigation information to the harvester host 10 based on the positioning information of the positioning device 40.
  • the image processing system 30 Based on the image of the farmland acquired by the image acquisition device 20, the image processing system 30 recognizes the unworked area 100, the operated area 200, and the field boundary area 300 from the image.
  • the image processing system 30 recognizes the unworked area 100, the worked area 200, and the field boundary area 300 from an image using image segmentation recognition technology. It can be understood that the image processing system 30 may also identify the area and boundary information in the image in other ways. Therefore, in the second preferred embodiment of the present invention, the manner in which the image processing system 30 recognizes the image is merely exemplary, not limiting.
  • the image processing system 30 recognizes the area of the farmland in the image, the boundary of the field, and the recognition based on the image around the harvester main body 10 taken by the image acquisition device 20 Information on the types of crops in the farmland, the height of the crops, the fullness of the grains, the thickness of the stems, etc.
  • the image processing system 30 is selected from any of the segmentation recognition methods selected from threshold-based segmentation methods, area-based segmentation methods, edge-based segmentation methods, and specific theory-based segmentation methods.
  • the image acquired by the acquiring device 20 performs segmentation recognition to identify regions and boundaries in the image.
  • the image processing system 30 uses a deep learning algorithm to recognize the image segmentation and perform area division and boundary definition on the image.
  • the image processing system 30 uses a deep learning algorithm to identify the area and boundary of the corresponding farmland in the image for the harvester host 10 to travel and perform operations based on the identified area and boundary.
  • the deep learning algorithm used by the image processing system 30 is a convolutional neural network algorithm image segmentation recognition technology to identify the unoperated area 100, the operated area 200, and the Described field boundary area 300.
  • the processing algorithm utilized by the image processing system 30 is merely exemplary, not limiting. Therefore, the image processing system 30 may also use other algorithms to perform segmentation recognition on the acquired image to identify the area and boundary of the farmland in the image.
  • the image processing system 30 is an image processor provided in the harvester host 10, wherein the image processor receives the image or image captured by the image acquisition device 20, and recognizes the Information in an image or video. According to the information recognized by the image processing system 30, the harvester host 10 correspondingly operates the parameters for controlling the driving path and adjusting the work.
  • the harvester host 10 further includes a vehicle body 11, an operating system 12 provided on the vehicle body 11, and a driving control system 13, wherein the operating system 12 is driven Connected to the vehicle body 11, the vehicle body 11 drives the working system 12 to drive the working system 12 to harvest crops.
  • the driving control system 13 controls the running of the vehicle body 11 and controls the operation of the working system 12. It is worth mentioning that the driving control system 13 has an unmanned driving mode and an operating driving mode. When the harvester is in the unmanned driving mode, the driving control system 13 controls the vehicle body 11 to automatically operate and the operation of the operation system 12. Accordingly, when the harvester is in the operation driving mode, the driving control system allows the driver to manually operate the vehicle body 11 and control the operation of the operation system by manual operation.
  • the driving control system 13 controls the driving of the vehicle body 11 and controls the harvesting operation of the working system 12. In other words, the driving control system 13 controls the adjustment of the operating parameters of the operating system 12 while the vehicle body 11 is traveling.
  • the driving control system 13 acquires the information of the image processing system 30 to identify the types of crops, the height of the crops, the degree of grain fullness, the diameter of the crop stalks, etc. in the image, and adjusts the operation based on the acquired information
  • the operating parameters of the system 12 are, for example, adjusting the operating speed of the operating system 12, the width of the operation, the height of the operation, and adjusting the parameters of post-processing.
  • the operating system 12 further includes at least one harvesting device 121, at least one conveying device 122, and at least one post-processing device 123, wherein the conveying device 122 is configured to receive the crops harvested by the harvesting device 121, and The crop is transported to the post-processing device 123 for the post-processing device 123 to post-process the crop.
  • the harvesting device 121, the conveying device 122, and the post-processing device 123 of the working system are respectively drivingly connected to the vehicle body 11, and the working system 12 is driven by the vehicle body 11 And operation of the harvesting device 121, the conveying device 122, and the post-processing device 123.
  • the post-processing device 123 is implemented as a post-processing subsequent processing device for crops, for example, a grain harvester, the post-processing device 123 is a threshing device, and the post-processing device 123 in the mowing equipment is Implemented as a packaging device, when the harvester is a vegetable and fruit harvesting device, the post-processing device 123 is implemented as a vegetable and fruit screening and storage device.
  • the driving control system 13 controls the width, height, and speed of the harvesting device 121 according to the image information recognized by the image processing system 30. It can be understood that when the density of crops in the farmland is large, the information of the crops in the farmland captured by the image acquisition device 20 is recognized by the image processing system 30, wherein the driving control system 13 processes the crops according to the image
  • the image information recognized by the system 30 controls any operating parameters such as reducing the harvesting width of the harvesting device 121, increasing the harvesting height, and reducing the harvesting speed.
  • the driving control system 13 controls the conveying speed, conveying power, etc. of the conveying device 122 according to the image information recognized by the image processing system 30. It can be understood that when the stalks of the crops in the farmland are thick, the height of the crops is high, and the density is large, the information of the crops in the farmland captured by the image acquisition device 20 is recognized by the image processing system 30, wherein the The driving control system 13 controls the operation parameters such as increasing the conveying speed of the conveying device 122 and increasing the conveying power according to the image information recognized by the image processing system 30.
  • the driving control system 13 controls the post-processing parameters of the post-processing device 123 according to the image information recognized by the image processing system 30. It can be understood that when the grains of the crops in the farmland are full, the size of the grains, the moisture content, the degree of dryness and wetness, and the types of crop fruits. It can be understood that the image processing system 30 recognizes the crop information of the crops in the farmland, wherein the driving control system 13 adjusts the post-processing according to the image information recognized by the image processing system 30
  • the post-processing parameters of the device such as the blowing power, the rotation speed of the post-processing chamber and other parameters.
  • FIG. 17 of the drawings of the specification of the present invention another alternative embodiment of an image acquisition device 20A of the harvester according to the second preferred embodiment of the present invention will be explained in the following description.
  • the image acquisition device 20A controls the angle of view and zoom of the lens inside the camera to prevent the camera from taking pictures.
  • the image acquisition device 20A includes a camera mounting mechanism 21A and at least one camera 22A, wherein the camera mounting mechanism 21A loads the camera 22A to the harvester host 10.
  • the bottom end of the camera mounting mechanism 21A is loaded to the harvester main body 10, and the camera mounting mechanism 21A is fixed by the harvester main body 10, wherein the upper end of the camera mounting mechanism 21A is set to be connected to the Camera 22A.
  • the camera 22A is supported by the camera mounting mechanism 21A to maintain relative balance so as to stably capture images or videos.
  • the camera 22A shoots images or videos around the harvester main body 10 under the support of the camera mounting mechanism 21A, wherein the camera 22A shoots the harvester based on the mounting position of the camera mounting mechanism 21A An image within the visual field of the host 10.
  • the camera 22A of the image acquisition device 20A acquires at least one visual image by photographing based on the position of the harvester host 10.
  • the camera 22A of the image acquisition device 20A acquires the image based on the field of view of the harvester host 10, thereby avoiding image data caused by changes in the position of the camera device 20A and the harvester host 10 Inaccurate question.
  • the present invention further provides an automatic driving method of a harvester, wherein the automatic driving method includes the following method steps:
  • the driving control system 13 controls the driving and operation of the harvester main body 10 based on the area information and the field boundary recognized by the image processing system 30.
  • Step (a) of the above-mentioned automatic driving method further includes: recognizing the information of the corresponding crops in the farmland in the image, wherein the information of the crops includes the types of crops, the height of the crops, the degree of grain fullness and the like.
  • Step (b) of the above-mentioned automatic driving method further includes steps:
  • (b.2) Plan at least one travel planning path 603 based on the identified area.
  • the step (b.1) of the above-mentioned automatic driving method further includes the steps of segmenting the image using an image segmentation technique, and identifying a region that divides the image.
  • step (a) of the above-mentioned automatic driving method based on the position and driving direction of the harvester main body 10, image information around the harvester main body 10 is photographed in real time.
  • the image acquisition device 20 captures an image near the position of the harvester main body 10 in real time.
  • step (b) of the above-mentioned automatic driving method the image processing system divides the image information using an image segmentation technique, and recognizes that the area dividing the image is the unoperated area 100, the operated area 200, ⁇ 300 ⁇ The field boundary area 300. Accordingly, the step (b.1) of the automatic driving method further includes the following steps:
  • the classification label of the image is output.
  • classification label corresponds to the unoperated area 100, the operated area 200, and the field boundary area 300.
  • the step (b.2) of the above-mentioned automatic driving method further includes the steps of: based on the positioning information of the harvester host 10, the image processing system 30 recognizes the area planning information of the image, and the navigation information of the navigation system 50 To get the driving plan path 603.
  • the step (b.2) of the above-mentioned automatic driving method further includes the step of adjusting the driving direction of the harvester body 10 based on the information of the image processing system 30 identifying the crops in the image to form a vehicle driving path 604.
  • the above-mentioned automatic driving method further includes: step (b.3) comparing whether the area division and the area boundary area identified by the image processing system 30 are consistent with the previous area boundary area, and if the consistency cannot be maintained, adjusting the image correspondence If the area division and area boundary range can be kept the same, the area division and boundary area will remain unchanged.
  • the driving control system 13 according to the positioning information of the harvester host 10, the regional planning information of the farmland obtained by the image processing system 30, and the navigation information To control the vehicle body 11 of the harvester host 10 to travel.
  • the automatic driving method further includes the step of: (d) adjusting the operating parameters of the operating system 12 of the harvester main machine 10 based on the identification information of the image.

Abstract

一种收割机及其自动驾驶方法，其中所述收割机包括一收割机主机(10)、至少一图像获取装置(20)、一路径规划***(60)以及一图像处理***(30)。所述图像获取装置(20)被设置于所述收割机主机(10)，所述图像获取装置(20)拍摄所述收割机主机(10)周围的图像，其中所述图像处理***(30)基于所述图像获取装置(20)拍摄的图像识别出所述图像中的农田信息，其中所述收割机主机(10)根据所述图像处理***(30)识别出的所述农田信息自动地控制驾驶，其中所述路径规划***(60)基于所述图像处理***(30)识别的所述农田信息规划出至少一行驶规划路径，其中所述收割机主机(10)根据所述路径规划***(60)规划出的所述行驶规划路径控制驾驶。

Description

收割机及其自动驾驶方法 技术领域
本发明涉及农业机械自动驾驶领域，尤其涉及一收割机及其自动驾驶方法。
背景技术
收割机是收获稻、麦等农作物作物子粒和秸秆的作物收获机械，此外收割机还包括割草机，收割其他农作物，比如蔬菜水果收获的机械设备。谷物收割机是一体化收割农作物的机械，能够一次性地完成收割、脱粒，并将谷粒集中到储藏仓，然后再通过传送带将粮食输送到运输车。果蔬类收获设备能够一次性地收割农田中的蔬菜水果，并将收获的果实与茎秆分离处理，然后分类处理。
收割机在执行收割作业时需要时刻观察农田中农作物和农田作业区域的情况，以便根据农作物的高度、成熟情况、颗粒饱满度等情况，调整所述收割机的作业参数。农业机械设备在农田中作业时，需要实时判断农田作业情况和农田中作物生长情况，来操作农业机械设备的运行和调整作业***的运行情况的复杂情况。现在只能通过具有专业技术的机械操作人员完成。农业机械设备在农田中作业时需要考虑到农田的已作业区域，未作业区域，以及天地的边界范围等诸多因素，并且在作业过程中需要根据农作物的情况实时地调整车辆的运行和调整作业的参数。由于驾驶过程中需要考虑复杂的作业环境，因此现有技术的农业设备还需要操作人员基于实时的农田作物的信息调整所述农业机械设备的运行。通过人工操作的方式控制所述农业机械设备作业出现判断误差的几率较大，而导致机械设备在作业过程中出现故障的概率大。
这种现有技术的收割机存在以下下述至少一缺陷：首先，收割机在作业时，由于车机本身的震动和农田土地的不平会产生所述收割机本体的上下晃动，从而导致设置于所述收割机本体的摄像头装置无法拍摄到稳定位置的图像。因此，通过所述摄像装置获取的图像往往是模糊不清的，无法为智能作业和自动驾驶提供信息支持。其次，现有技术的摄像装置通过固定安装的方式设置于所述收割机本体，仅仅能够获取单一方向的图像，比如所述收割机前方的图像，而不能根据情况调整所述摄像装置的拍摄方向和位置。再次，现有技术的移动摄像设备或固定摄像设备，比如无人机的摄像装置或固定在农田中的摄像装置，拍摄所述收割机 周围的图像后传输至所述收割机本体，以供所述收割机本体读取所述摄像装置拍摄的图像。虽然在一定程度上解决了图像拍摄不清楚的问题，但是，基于摄像装置本身或基于无人机的位置拍摄的图像，不能从所述收割机自身的视角获取图像。因此，获取的图像不能很好的识别。
现有技术的农业机械设备在作业时通常会由于设定的作业路径不准，而导致作业出现误差，甚至出现严重的机械故障。此外，采用PTK的卫星定位的方式对于农业设备的性能要求较高，所需要的制造成本和维护成本相对来说都很高昂，因此，这种现有技术的自动驾驶的定位方式不适用于当前的农业机械设备的自动驾驶模式中。
发明内容
本发明的一个主要优势在于提供一收割机及其自动驾驶方法，其中所述收割机基于拍摄的至少一视觉图像识别出所述图形中农田的区域。
本发明的另一个优势在于提供一收割机及其自动驾驶方法，其中所述收割机基于拍摄的至少一视觉图像识别出所述图形中作物的种类、高度、成熟情况等作物信息。
本发明的另一个优势在于提供一收割机及其自动驾驶方法，其中所述收割机基于所述视觉图像识别出所述视觉图像中的未作业区域、已作业区域、以及田边界区域，以便根据识别出的所述区域控制收割机的行驶路径。
本发明的另一个优势在于提供一收割机及其自动驾驶方法，其中所述收割机基于所述视觉图像识别出所述图像中作物的信息，其中所述收割机根据所述图像中的识别信息调整收割机的作业参数，提高所述收割机的作业质量、效率。
本发明的另一个优势在于提供一收割机及其自动驾驶方法，其中所述收割机的图像获取装置为云台摄像装置，其中所述云台摄像装置具有防抖拍摄功能，提高了所述收割机获取视觉图像的准确性、稳定性。
本发明的另一个优势在于提供一收割机及其自动驾驶方法，其中所述图像获取装置被设置于所述收割机的一收割机主体，其中所述收割机通过所述图像获取装置拍摄所述收割机主体周围的图像。
本发明的另一个优势在于提供一收割机及其自动驾驶方法，其中所述图像获取装置被设置于所述收割机的一收割机主体，其中所述图像获取装置被设置于所 述收割机主体，其中所述图像获取装置基于所述收割机主体的视野位置拍摄至少一视觉图像或视觉影像，以便根据拍摄的影像信息识别所述收割机主体周围的信息。
本发明的另一个优势在于提供一带有收割机及其自动驾驶方法，其中所述图像获取装置可基于所述收割机主机的位置，被调整地拍摄不同角度和不同方向的图像，以便于获取所述收割机主机不同方向的图像。
本发明的另一个优势在于提供一带有收割机及其自动驾驶方法，其中所述图像获取装置为一机械云台相机或电子云台相机，藉由所述图像获取装置提高所述视觉图像的稳定性。
本发明的另一个优势在于提供一收割机及其自动驾驶方法，其中所述收割机的一路径规划***基于当前车辆的定位信息、图像处理***识别的信息以及导航***的信息自动地规划路径。
本发明的另一个优势在于提供一收割机及其自动驾驶方法，其中所述收割机基于所述视觉图像识别出的所述区域，规划出所述收割机的行驶路径和作业路线。
本发明的另一个优势在于提供一收割机及其自动驾驶方法，其中所述收割机的一图像获取装置实时地获取周围农田的所述视觉图像，实时地更新所述收割机规划的路径导航信息。
本发明的另一个优势在于提供一收割机及其自动驾驶方法，其中所述收割机通过所述图像获取装置实时拍摄地图像，识别出所述视觉图像中的所述区域，并且根据区域的变化实时地更新或调整所述收割机的作业路线，提高所述收割机的作业质量。
本发明的另一个优势在于提供一收割机及其自动驾驶方法，其中所述收割机的所述图像处理***基于获取的所述视觉图像信息，利用图像分割技术识别出图像中的所述未作业区域、所述已作业区域、以及所述田边界区域，以及划分相邻两区域的边界。
本发明的另一个优势在于提供一收割机及其自动驾驶方法，其中所述收割机的所述图像处理***基于获取的所述视觉图像信息，利用图像分割技术识别出图像中农作物的种类、高度、颗粒饱满度等植物信息，以供所述收割机的作业***基于农作物的信息调整作业参数。
本发明的另一个优势在于提供一收割机及其自动驾驶方法，其中所述收割机 的所述图像处理***基于获取的图像信息识别出所述图像中的区域边界，以便所述路径规划***基于识别出的所述区域边界规划车辆的行驶路径。
本发明的另一个优势在于提供一收割机及其自动驾驶方法，其中所述收割机不需要高精度的卫星定位，降低了所述自动驾驶设备的生产制造难度，同时也降低了设备的维护成本。
本发明的另一个优势在于提供一收割机及其自动驾驶方法，其中所述收割机基于所述图像处理***输出的区域划分信息，进行路径规划，以实现自动驾驶和自动驾驶作业。
本发明的其它优势和特点通过下述的详细说明得以充分体现并可通过所附权利要求中特地指出的手段和装置的组合得以实现。
依本发明的一个方面，能够实现前述目的和其他目的和优势的本发明的一收割机，包括：
一收割机主机；
至少一图像获取装置，其中所述图像获取装置被设置于所述收割机主机，所述图像获取装置拍摄所述收割机主机周围的图像，以及
一图像处理***，其中所述图像处理***基于所述图像获取装置拍摄的图像识别出所述影像中的农田信息，其中所述收割机主机根据所述图像处理***识别出的所述农田信息自动地控制驾驶。
根据本发明的一实施例，所述收割机进一步包括一路径规划***，其中所述路径规划***基于所述图像处理***识别的所述农田信息规划出至少一行驶规划路径，其中所述收割机主机根据所述路径规划***规划出的所述行驶规划路径控制驾驶。
根据本发明的一实施例，所述图像处理***利用图像分割识别技术识别出所述图像中农田的信息，和基于识别出的信息规划所述图像中农田的区域。
根据本发明的一实施例，所述图像处理***利用图像分割识别技术识别出所述图像中农作物信息，以供所述收割机主机基于识别出的信息自动地调整作业参数。
根据本发明的一实施例，所述图像获取装置为防抖云台摄像装置，所述图像获取装置被装载于所述收割机主机，基于所述收割机主机的位置以拍照的方式拍摄所述收割机主机周围的图像。
根据本发明的一实施例，所述图像获取装置为机械防抖云台装置，所述图像获取装置包括一云台和至少一摄像机，其中所述云台将所述摄像机安装至所述收割机主机，所述摄像机被设置于所述云台，藉由所述云台支撑所述摄像机保持平衡。
根据本发明的一实施例，所述图像获取装置为电子云台装置，所述图像获取装置通过控制镜头的视角和变焦，从而防止所述图像获取装置镜头拍照抖动。
根据本发明的一实施例，所述图像获取装置被设置于所述收割机主机的前部、所述收割机主机的顶部、所述收割机主机的左侧、右侧、或所述收割机主机的后部。
根据本发明的一实施例，所述图像处理***进一步包括：
一图像分割模块，其中所述图像分割模块分割所述图像为多个像元区域，其中每一所述像元区域包括至少一像素单元；
一特征化模块，其中所述特征化模块基于所述像元区域的所述像素单元提取每一像元区域对应的特征；以及
一区域划分模块，其中所述区域规划模块根据所述像元区域的特征识别和划分所述图像的区域。
根据本发明的一实施例，所述收割机进一步包括一定位装置和一导航***，所述定位装置和所述导航***被设置于所述收割机主机，其中所述定位装置获取所述收割机主机的位置信息，其中所述导航***为所述谷物处理主体提供导航信息。
根据本发明的一实施例，所述路径规划***进一步包括：
一作业区域设置模块，其中所述作业区域设置模块设定所述农田的边界区域得到的所述农田的作业区域和所述作业边界；和
一行驶路径规划模块，其中所述基于所述收割机主机的定位信息，所述图像处理***识别所述图像的区域规划信息，以及所述导航***的导航信息，得出至少一行驶规划路径。
根据本发明的一实施例，所述收割机主机包括一车辆主体，设置于所述车辆主体的至少一作业***，以及一驾驶控制***，所述车辆主体驱动所述作业***运行，其中所述驾驶控制***控制所述车辆主体的运行和控制所述作业***的作业参数。
根据本发明的一实施例，所述驾驶控制***获取所述图像处理***识别的所述图像获取装置拍摄的图像的信息，自动地控制所述车辆主体的行驶路线和控制所述作业***的作业参数，以实现无人自动驾驶和收割作业。
根据本发明的另一方面，本发明进一步提供一收割机的自动驾驶方法，其中所述自动驾驶方法包括如下步骤：
(a)获取至少一图像，和识别所述图像中农田的区域和田边界；
(b)基于所述识别信息，规划出至少一行驶规划路径；以及
(c)控制所述收割机主机按照所述行驶规划路径自动地行驶。
根据本发明的一实施例，上述自动驾驶方法的步骤(a)进一步包括：识别出所述图像中对应农田中农作物的信息，其中所述农作物的信息包括农作物种类，农作物的高度，颗粒饱满度等信息。
根据本发明的一实施例，上述自动驾驶方法步骤(b)进一步包括步骤：
(b.1)识别划分出所述图像对应农田的区域和边界；以及
(b.2)基于识别的所述区域规划出至少一行驶规划路径。
根据本发明的一实施例，上述自动驾驶方法的步骤(b.1)进一步包括步骤：利用图像分割技术分割所述图像，和识别划分所述图像的区域。
根据本发明的一实施例，在上述自动驾驶方法的步骤(b)中，所述图像处理***利用图像分割技术分割所述图像信息，和识别划分所述图像的区域为所述未作业区域、所述已作业区域、以及所述田边界区域。
根据本发明的一实施例，所述自动驾驶方法的步骤(b.1)进一步包括如下步骤：
分割所述图像为多个所述像元区域，和归一化所述像元区域的像素值为一数组；
提取每一数组对应的所述像元区域的特征；以及
基于所述像元区域对应的特征，输出所述图像的分类标签。
根据本发明的一实施例，上述自动驾驶方法的步骤(b.2)进一步包括步骤：基于所述收割机主机的定位信息，所述图像的区域规划信息，以及所述导航***的导航信息，规划出所述行驶规划路径。
根据本发明的一实施例，上述自动驾驶方法进一步包括：步骤(b.3)对比所述图像处理***识别出的区域划分和区域边界范围与之前的区域边界范围是 否保持一致，若不能保持一致，则调整所述图像对应的区域划分和区域边界范围，若能够保持一致，则保持区域划分和边界范围不变。
根据本发明的一实施例，上述自动驾驶方法进一步包括步骤：(d)基于所述图像的识别信息，调整所述收割机主机的作业***的作业参数。
通过对随后的描述和附图的理解，本发明进一步的目的和优势将得以充分体现。
本发明的这些和其它目的、特点和优势，通过下述的详细说明，附图和权利要求得以充分体现。
附图说明
图1是根据本发明的第一较佳实施例的一收割机的***示意图。
图2是根据本发明的上述较佳实施例的所述收割机的图像获取示意图。
图3A是根据本发明的上述较佳实施例的所述收割机获取的一种图像的示意图。
图3B是根据本发明的上述较佳实施例的所述收割机获取的另一种图像的示意图。
图3C是根据本发明的上述较佳实施例的所述收割机获取的另一种图像的示意图。
图4是根据本发明的上述较佳实施例的所述收割机的一图像处理***划分识别所述图像区域的示意图。
图5A是根据本发明的上述较佳实施例的所述收割机的所述图像处理***分割所述图像区域的示意图。
图5B是根据本发明的上述较佳实施例的所述收割机的所述图像处理***的***框图。
图6是根据本发明的上述较佳实施例的所述收割机的所述图像处理***提取所述图像区域特征识别的示意图。
图7是根据本发明的上述较佳实施例的所述收割机的所述图像处理***输出所述图像的区域划分示意图。
图8是根据本发明的上述较佳实施例的所述收割机的所述图像处理***输出所述图像的区域划分的界线划分变化示意图。
图9是根据本发明的上述较佳实施例的所述收割机的自动驾驶场景示意图。
图10是根据本发明的上述较佳实施例的所述收割机的一路径规划***的***示意图。
图11A是根据本发明的上述较佳实施例的所述收割机的所述路径规划***生成的农田路径规划示意图。
图11B是根据本发明的上述较佳实施例的所述收割机的所述路径规划***调整行驶路径的示意图。
图12是根据本发明的第二较佳实施例的一带有图像获取装置的收割机的整体结构的示意图。
图13是根据本发明的上述较佳实施例的所述收割机的所述图像获取装置拍摄图像的示意图。
图14是根据本发明的上述较佳实施例的所述收割机的所述图像获取装置的结构示意图，其中所述图像获取装置被实施为一机械云台装置。
图15是根据本发明的上述较佳实施例的所述收割机的所述图像获取装置被安装位置的示意图。
图16A是根据本发明的上述较佳实施例的所述收割机的一图像处理***识别所述图像获取装置拍摄图像中农田区域的示意图。
图16B是根据本发明的上述较佳实施例的所述收割机的一图像处理***识别所述图像获取装置拍摄图像中农作物的示意图。
图17是根据本发明的上述较佳实施例的所述收割机的所述图像获取装置的另一可选实施方式的示意图，其中所述图像获取装置被实施为一电子云台装置。
具体实施方式
以下描述用于揭露本发明以使本领域技术人员能够实现本发明。以下描述中的优选实施例只作为举例，本领域技术人员可以想到其他显而易见的变型。在以下描述中界定的本发明的基本原理可以应用于其他实施方案、变形方案、改进方案、等同方案以及没有背离本发明的精神和范围的其他技术方案。
本领域技术人员应理解的是，在本发明的揭露中，术语“纵向”、“横向”、“上”、“下”、“前”、“后”、“左”、“右”、“竖直”、“水平”、“顶”、“底”、“内”、“外”等指示的方位或位置关系是基于附图所示的方位或位置 关系，其仅是为了便于描述本发明和简化描述，而不是指示或暗示所指的装置或元件必须具有特定的方位、以特定的方位构造和操作，因此上述术语不能理解为对本发明的限制。
可以理解的是，术语“一”应理解为“至少一”或“一个或多个”，即在一个实施例中，一个元件的数量可以为一个，而在另外的实施例中，该元件的数量可以为多个，术语“一”不能理解为对数量的限制。
参照本发明说明说明书附图之图1至图9，依照本发明第一较佳实施例的一收割机及其自动驾驶方法在接下来的描述中被揭露和被阐述，其中所述收割机可被实施为具有谷物处理功能的农作物收割机设备、蔬菜、水果收获设备、割草设备、以及其他类型的收割装置等。可以理解的是，本发明中所述的收割机的类型在此仅仅作为示例性质的，而非限制。所述收割机获取周围的至少一图像，和基于视觉地识别处理所述图像中的农田的区域类型，并划分所述图像中的农田的各种区域类型和边界。
所述收割机根据划分的各个区域类型和边界，其中所述收割机划分农田的区域类型包括至少一已作业区域100、至少一未作业区域200、以及至少一田边界区域300，并且所述收割机根据划分的区域类型由导航***规划出车辆的行走路线，以实现无人自动驾驶和无人自动驾驶地作业。
值得一提的是，自动驾驶汽车，在自动驾驶模式下为识别获取车辆行走的路线，需要获取精确的车辆定位信息，通常需要高精度的卫星定位信息，并且所述自动驾驶汽车需时刻更新道路中存在的障碍物信息、路面车辆信息、以及路面行人等信息，以在高速运行状态下实现自动驾驶功能。本发明的所述收割机获取的图像是对应于农田中的农作物谷物的图像数据信息，其中所述图像是基于车辆当前位置获取的所述车辆周边的图像。所述收割机不需要太高精度的卫星定位信息，只需要普通米级精度的卫星定位(GPS定位或者北斗定位等)。相应地，所述收割机所获取和处理的图像与自动驾驶汽车不同，因此，所述收割机所形成的路径规划和驾驶方式也不相同。可以理解的是，本发明的所述收割机基于视觉的识别所述农田的区域和自动驾驶功能与自动驾驶汽车的识别模式不同。
如图1和图2所示，所述收割机获取周边的至少一图像，其中所述收割机根据获取得到的所述图像识别划分出所述农田对应的区域类型和区域之间的界线。所述收割机通过定点拍照、摄像，移动地拍照等方式获取所述收割机周边的所述 图像。可以理解的是，所述收割机获取图像的方式在此仅仅作为示例性质的，而非限制。所述收割机包括一收割机主机10和至少一图像获取装置20，其中所述图像获取装置20获取所述收割机主机10周边的至少一图像。
优选地，所述图像获取装置20被设置于所述收割机主机10，其中所述图像获取装置20通过拍照或视频拍摄的方式获取所述收割机主机10周边的所述图像。更优选地，所述图像获取装置20被设置于所述收割机主机10的前方，其中所述图像获取装置20能够实时地获取所述收割机主机10前方的图像，其中所述收割机主机10基于所述图像获取装置20拍摄的图像信息所识别划分出的区域设定行驶路线。值得一提的是，所述图像获取装置20拍摄的图像是基于所述收割机主机10视野范围内的图像。换言之，所述图像获取装置20获取基于收割机主机10视野方向的图像，根据所述图像获取装置20被安装至所述收割机主机10的位置调整所述收割机主机10的行驶方向。
更多地，所述图像获取装置20拍摄所述收割机主机10行驶方向的天地的视觉，其中所述图像可以是被拍摄得到的二维平面图像或三维立体图像。可以理解的是，所述图像获取装置20拍摄得到的图像的类型在此仅仅作为示例性质的，而非限制。
值得一提的是，在本发明的第一较佳实施例中，所述收割机主机10被实施为一谷物收割机设备，其中所述收割机主机10被控制地行驶至农田的未作业区域200进行收割作业，以收割所述未作业区域200内的农作物，比如水稻，小麦，玉米等。所述收割机主机10根据所述图像获取装置20获取的图像划分出的区域进行田间的自动驾驶，无人自动驾驶。可以理解的是，所述收割机主机10的类型在此仅仅作为示例性质的，而非限制。
如图3A至图3C所示，所述收割机主机10在行驶过程中，由所述图像获取装置20实时地获取所述收割机主机10周边的图像。图3A示出了所述收割机主机10作为一谷物收割机时，所述图像获取装置20拍摄的所述图像。农田中的区域根据所述谷物是否收获情况分为至少一未收割区域100a，至少一已收割区域200a，以及至少一田边界区域300a，其中所述已收割区域200a是已经收割农作物的区域，其中所述已收割区域200a中原有的农作物被收割。所述未收割区域100a是还农作物还存在的区域，其中所述未收割区域100a中还存在生长的农作物。所述田边界区域300a是农田中分隔农作物间隔的地垄、农田周边的外边界、 以及农田中存在障碍物区域，其中所述田边界区域300a不被种植农作物。
图3B示出了所述收割机主机10作为一割草设备时，所述图像获取装置20拍摄的所述图像。农田中的区域根据所述谷物是否耕种分为至少一未收割区域100b，至少一已收割区域200b，以及至少一田边界区域300b，其中所述未收割区域100b代表还未收割农作物的区域，所述已收割区域200b代表已经被耕种农作物的区域，所述田边界300b是间隔农作物种植的地垄和农田周边的外边界、以及农田中存在障碍物区域。
图3C示出了所述收割机主机10作为一根茎类植物或水果收获设备时，比如蔬菜收割机设备，所述图像获取装置20拍摄的所述图像。农田中的区域根据所述谷物是够已经喷药分为至少一未收获区域100c，至少一以收获区域200c，以及至少一田边界区域300c。所述未收获区域100c代表还未农作物还未收获的区域，所述以收获区域200c代表已经被收获的农作物区域，所述田边界300b是间隔农作物种植的地垄和农田周边的外边界、以及农田中存在障碍物区域。
如图1和图4所示，所述图像获取装置20获取的图像被利用图像分割识别技术识别出所述未作业区域100、所述已作业区域200、以及所述田边界区域300，并区分所述区域之间的界线。所述收割机进一步包括一图像处理***30，其中所述图像处理***30基于所述图像获取装置20获取的所述农田的所述图像，利用图像分割识别技术从图像中识别出所述未作业区域100、所述已作业区域200、以及所述田边界区域300。
可以理解的是，所述图像处理***30利用图像分割识别技术识别所述图像中的区域和边界被用于表示所述收割机主机10行驶前方的所述农田的区域和边界。基于所述图像处理***30利用图像分割识别技术识别出的所述区域和边界，所述收割机主机10被控制地在农田中的未作业区域行驶和进行作业。比如收割机设备，被设置于所述收割机设备前端的所述图像获取装置20获取收割机设备前方的农田的图像，其中所述图像获取装置20拍摄的图像被所述图像处理***30分割识别，以识别划分出所述收割机设备行驶方向的农田的所述未作业区域100、所述已作业区域200、以及所述田边界区域300。所述收割机主机10即收割机设备的主机基于所述图像处理***30识别出的区域和边界规划车辆行驶路径和收割作业。
可以理解的是，所述图像处理***30利用图像分割识别技术识别所述图像 获取装置20提供的图像中农作物的种类，农作物的高度、农作物颗粒饱满程度等信息。所述图像处理***30可基于识别出的所述图像中农作物的种类，农作物的高度判断农作物是否已经收割，基于识别出的所述图像中农作物颗粒的饱满程度信息，可被用于调整作业的参数。换言之，所述图像处理***30可根据所述图像获取装置20提供的图像识别出农田的区域类型和边界，还可以识别出农田中农作物的种类、高度、颗粒饱满度、农作物成熟情况等。
值得一提的是，所述图像处理***30选自基于阈值的分割方法、基于区域的分割方法、基于边缘的分割方法以及基于特定理论的分割方法等其中的任一分割识别方法对所述图像获取装置20获取的图像进行分割识别，以识别出所述图像中的区域和边界。优选地，所述图像处理***30利用深度学习算法对所述图像分割识别和对所述图像进行区域划分和边界的限定。换言之，所述图像处理***30利用深度学习算法识别所述图像中对应的农田的区域和边界，以供所述收割机主机根据识别划分的区域和边界行驶和进行作业。可更优选地，所述图像处理***30利用的深度学习算法为卷积神经网络算法的图像分割识别技术从图像中识别出对应农田中的所述未作业区域100、已作业区域200、以及所述田边界区域300。
值得一体地是，所述图像处理***30利用的处理算法在此仅仅作为示例性质的，而非限制。因此，所述图像处理***30还可利用其它算法对获取的图像进行分割识别，以识别出图像中农田的区域和边界。
如图5A和图6所示，所述图像处理***30对所述图像获取装置20获取的所述图像分割成为多个像元区域301，其中每一所述像元区域301内包含至少一像素单元。可以理解的是，所述图像对应于所述收割机主机10周围的区域，相应地，所述图像的所述像元区域301对应于被拍摄的农田中的特定区域的农田或农作物的图像信息。对分割形成的每一所述像元区域301进行归一化处理，使得所述像元区域301的所述像素单元归一化为与像素值对应大小的一数值或数组。换言之，所述图像处理***30将分割形成的所述像元区域301归一化为对应的数值或数组，以供所述图像处理***提取图像的特征，和对区域的划分。
所述图像处理***30基于每一所述像元区域301所对应的数组，提取所述像元区域301对应的图像特征。所述图像处理***30根据所述像元区域301对应的所述数组得出所述像元区域301对应的图像特征。所述图像处理***30利 用卷积神经网络算法，比如二维卷积神经网络时，所述卷积神经网络的输入层对应输入所述像元区域301中对应的二维数组或三维数组。所述卷积神经网络的隐含层对输入层的数组进行特征提取，和在特征提取后进行特征选择和信息过滤。所述卷积神经网络基于所述数组对应的特征输出所述像元区域301的一分类标签，其中所述分类标签分别对应所述未作业区域100、已作业区域200、以及所述田边界区域300。
如图6和图7所示，所述图像处理***30通过提取所述像元区域301对应数组的特征，识别所述像元区域301对应的区域特征，其中所述像元区域301对应的特征主要包括有农作物植株的高度特征、农田中农作物植株的间隔、农作物颜色、农田土地颜色、农作物种类特征、农田土地特征、农作物颗粒饱满程度、农作物颗粒数量等。所述图像处理***30根据提取特征输出对应像元区域301的一分类标签，其中所述分类标签基于所述特征信息对应地标识出所述像元区域301对应的区域类型和边界界线。
如图5B所示，所述图像处理***30包括一图像分割模块31、一特征化模块32、以及一区域划分模块33。所述图像分割模块31获取所述图像获取模块20拍摄的图像，和对图像分割处理形成多个所述像元区域301，其中每一所述像元区域301对应包括至少一像素单元。所述特征化模块32利用深度学习算法提取所述像元区域301对应的特征类型，和对特征选择，以及信息过滤。所述区域划分模块33基于所述特征化模块32提取的所述像元区域301对应的特征，对所述图像划分，以生成对应所述未作业区域100、已作业区域200、以及所述田边界区域300的分类标签。
优选地，所述图像分割模块31分割所述图像为多个所述像元区域301，其中每一所述像元区域301的大小、形状和范围相同。可以理解的是，所述图像分割模块31还可根据所述图像像素阈值大小进行分割，换言之，所述退昂分割模块31分割的所述像元区域301的大小、形状和范围可不同。更优选地，所述图像处理***30的所述特征化模块32采用卷积神经网络算法时，所述图像分割模块31分割的所述像元区域301为单个像素单元。
所述特征化模块32包括一像元处理模块321、一特征提取模块322、以及一特征输出模块323，其中所述像元处理模块321处理所述像元区域301中像素单元对应的数组。换言之，所述像元处理模块321将所述像元区域301归一化为适 于处理的数组。所述特征提取模块322输入所述像元处理模块321处理的所述像元区域301的数组后提取所述数组对应的特征类型，和对特征进行选择，以及信息过滤，以保留可用数据，排出干扰数据，进而使得特征提取结果更准备。所述特征输出模块323输出由所述特征提取模块322结合所述特征提取模块322提取的特征，藉由所述区域划分模块33结合所述特征输出模块323输出的特征，生成对应区域的所述分类标签。
所述区域划分模块33基于所述特征化模块32提取得到的所述像元区域301对应的特征，划分所述图像对应的各个区域，和设定区域边界。相应地，所述区域划分模块33进一步包括一区域分割模块331和一边界划分模块332，其中所述区域分割模块331根据所述像元区域301的特征划分不同的区域，其中所述边界划分模块332划分所述区域对应的边界范围，以便于认定区域的范围。
所述收割机的所述收割机主机10在行驶过程中，由所述图像获取装置20实时地获取所述收割机主机10前方视野范围内的图像。相应地，所述图像处理***30实时地获取所述图像获取装置20拍摄的图像，和利用图像分割识别技术识别出所述图像对应于农田的区域划分和区域边界范围。当所述图像处理***30识别出的区域划分和区域边界范围不能与之前的区域边界范围保持一致时，调整所述图像对应的区域划分和区域边界范围。
如图8所示，所述收割机主机10在行驶作业过程中，不可避免的会产生振动，和行驶方向偏移等问题。当所述收割机主机10行驶方向偏移，或因为车辆的震动而造成的区域划分变化时，所述图像处理***30实时地更新所述图像对应的区域划分和区域边界范围。
如图1所示，所述收割机进一步还包括一定位装置40和一导航***50，其中所述定位装置40被设置于所述收割机主机10，以获取所述收割机主机10的位置定位信息。优选地，所述定位装置40利用卫星定位信息获取所述收割机主机10的位置信息，比如GPS或北斗定位装置等。所述导航***50被设置于所述收割机主机10，其中所述导航***50对所述收割机主机10的行驶进行导航，以供所述收割机主机10基于利用所述定位装置40的定位信息和所述图像处理***30得到的区域规划信息，以及所述导航***50的导航信息，实现无人自动驾驶和作业。
可以理解的是，所述图像处理***30基于所述图像而得到的农田的区域划 分和区域边界范围信息等被实时地更新至所述导航***50，以更新所述导航***50的导航信息。优选地，所述导航***50被实施为惯性组合导航***。可以理解的是，所述导航***50的类型在此仅仅作为示例性质的，而非限制，因此，所述导航***50还可以被实施为其他类型的导航装置。
相应地，所述收割机的所述收割机主机10包括一车辆主体11，设置于所述车辆主体11的一作业***12，以及一驾驶控制***13，其中所述作业***12被所述车辆主体11带动，和实现谷物处理的作业，比如收割作业。所述驾驶控制***13控制所述车辆主体11的行驶和控制所述作业***12的作业。值得一提的是，所述驾驶控制***13具有一无人驾驶模式和一操作驾驶模式。当所述收割机处于所述无人驾驶模式时，所述驾驶控制***13控制所述车辆主体11自动地运行和所述作业***12的作业。相应地，当收割机处于所述操作驾驶模式时，所述驾驶控制***允许驾驶人员通过人工操作的方式操作所述车辆主体11的运行和控制所述作业***的作业。
在本发明的第一较佳实施例中，所述收割机为收割机设备，其中所述作业***12被实施为一收割作业设备。所述驾驶控制***13控制所述车辆主体11的行驶和控制所述作业***12的作业。换言之，所述驾驶控制***13控制所述车辆主体11在行驶的过程中所述作业***12作业参数的调整。所述驾驶控制***13获取所述图像处理***30识别所述图像中的农作物的种类、农作物高度、颗粒饱满程度、农作物茎秆的直径大小等信息，和基于获取的所述信息调整所述作业***12的作业参数，比如，调整所述作业***12作业速度，作业的宽幅，作业的高度，调整脱力处理的参数等。
本发明说明书附图之图9示出了所述收割机在农田中无人驾驶和收割作业的实施方式。所述收割机主机10的所述驾驶控制***13处于所述无人驾驶模式时，所述驾驶控制***13获取所述定位装置40提供的所述车辆主体11的定位信息、所述导航***50提供的导航信息、以及所述图像处理***30提供的区域识别信息，进而控制所述车辆主体11行驶在所述农田的所述未作业区域100，以完成谷物的收割作业。所述收割机主机10在行驶作业过程中，所述图像获取装置20实时地获取所述车辆主体11行驶前方的所述图像，其中所述图像被所述图像处理***30利用图像分割识别技术识别出区域范围和边界范围。当所述图像处理***30得到的区域划分和边界范围与之前的区域划分和边界范围不一致时，所 述图像处理***30替换原有的区域划分和边界范围，和更新所述导航***50的导航数据，以使所述驾驶控制***13获取新的导航信息调整行驶和作业路线。
如图10至图11B所示，所述收割机基于所述定位装置40获取的所述收割机主机10的位置信息，所述图像处理***30识别的所述图像的区域规划信息，以及所述导航***50的导航信息，生成至少一规划路径。所述收割机主机10的所述驾驶控制***13根据生成的所述规划路径控制所述车辆主体11的行驶和控制所述作业***12的作业。相应地，所述收割机进一步包括一路径规划***60，其中所述路径规划***为所述收割机主机10规划至少一车辆的行驶路径。所述路径规划***60获取所述定位装置40的定位信息，获取所述图像处理***30识别的所述图像的区域规划信息，以及获取所述导航***50的导航信息，和根据获取的信息规划所述车辆主体11的行驶路径。
如图11A和图11B所示，所述路径规划***60识别或者设定出所述农田中对应的至少一作业区域601和作业边界602，其中所述作业区域601是所述收割机最大的作业范围，其中所述驾驶控制***13控制所述车辆主体11行驶在所述作业边界602的范围内。可以理解的是，所述作业区域601和所述作业边界602可由所述图像处理***30通过识别图像中的所述田边界区域300的方式识别出所述作业区域601的最大区域范围和边界。或者，所述路径规划***60通过设定的方式设定出所述收割机的所述作业区域601。
所述路径规划***60基于所述作业区域601的最外侧的所述作业边界602规划出至少一行驶路径。当所述作业区域601的宽度大于所述作业***12的作业宽度时，所述路径规划***60规划出“回”字型的行驶路线，或“S”型的行驶路线。可以理解的是，所述路径规划***60规划出的行驶路线的方式在此仅仅作为示例性质的，而非限制。因此，其他方式的行驶路线也可应用于此。
优选地，当所述车辆主体11行驶至所述作业区域601的远端边界时，所述路径规划***60基于当前未作业区域100的范围重新规划出至少一行驶路径。换言之，当所述车辆主体11行驶至所述作业区域601的远端边界时，由所述路径规划***60为所述车辆主体11更新所述作业区域601和所述作业边界602，并根据更新的所述作业区域601规划出新的行驶路径。
可以理解的是，所述驾驶控制***13控制所述车辆主体11按照所述路径规划***60规划出的行驶路径行驶，其中所述驾驶控制***13控制所述作业*** 12收割所述作业区域401最外侧的农作物。换言之，所述驾驶控制***13控制所述作业***12基于所述作业边界602收割所述未作业区域100内的作物。
如图10所述，所述收割机的所述路径规划***60包括一作业区域设置模块61、一行驶路径规划模块62、以及一路径调整模块63。所述作业区域设置模块61基于所述图像处理***30识别所述图像中所述农田的边界区域得到的所述农田的作业区域601和所述作业边界602；或者通过设置的方式设置所述收割机主机10作业在所述农田中的所述作业区域601和所述作业边界602。由于所述收割机主机10的作业使得未作业区域100和所述已作业区域200变化，所述作业区域设置模块61实时地更新所述作业区域601的范围和所述作业边界602的界线，以便生成新的所述未作业区域100，和所述已作业区域200。
所述行驶路径规划模块62基于所述收割机主机10的定位信息，所述图像处理***30识别所述图像的区域规划信息，以及所述导航***50的导航信息，得出至少一行驶规划路径603，其中所述驾驶控制***13控制所述车辆主体11按照所述行驶规划路径603行驶。所述路径调整模块63基于所述图像处理***30识别所述图像的农作物的信息调整所述收割机主体10的行驶方向，以形成一车辆行驶路径604，其中所述车辆行驶路径604基本重合或平行于所述行驶规划路径603。当所述图像处理***30识别图像中农作物需要调整收割范围时，所述路径调整模块63生成的所述车辆行驶路径偏离于所述行驶规划路径603。
参照本发明说明书附图之图12至图16B所示，依照本发明第二较佳实施例的带有图像获取装置的一收割机在接下来的描述中被阐明。所述收割机包括一收割机主机10和至少一图像获取装置20，其中所述图像获取装置20被设置于所述收割机主机10，所述图像获取装置20拍摄所述收割机主机10所在农田的图像或视频影像，以供所述收割机主机10基于所述图像获取装置20拍摄的图像或影像信息控制行驶方向和/或作业参数。所述图像获取装置20基于所述收割机主机10的位置拍摄所述收割机主机10所在农田位置周围的农田的信息。可以理解的是，所述图像获取装置20捕获视野范围内的图像，比如驾驶人员视野中的图像，以便根据拍摄到的所述图像调整所述收割机主机10的运行参数，比如调整行驶路线，行驶速度、作业参数等。
值得一提的是，所述图像获取装置20被搭载至所述收割机主机10，其中所述图像获取装置20捕获到的图像和影像信息被传输至所述收割机主机10，以供 所述收割机主机10基于所述信息调整运行参数。所述图像获取装置20被搭载至所述收割机主机10，其中所述图像获取装置20在收割机主机10产生抖动时拍摄出清晰影像。换言之，所述图像获取装置20为防抖摄像装置，能够在拍摄时避免所述收割机主机10自身的机械振动和由于天地不平造成的抖动。所述收割机主机10基于所述图像获取装置20拍摄的图像信息，在操作人员的操作下或者由自动地控制行驶路径和作业参数，以实现所述收割机的运行作业。换言之，所述收割机主机10基于所述图像获取装置20拍摄的图像信息调整运行和作业参数，以实现精准地作业/或无人自动驾驶作业。
优选地，在本发明的第二较佳实施例中，所述图像获取装置20被实施为云台摄像装置，其中所述图像获取装置20在震动或抖动的情况下，拍摄出稳定性质的图像或影像。
在本发明的第二较佳实施例中，所述图像获取装置20为机械云台装置，其中所述图像获取装置20通过机械连接的方式被搭载至所述收割机主机10，并且所述图像获取装置20通过机械防抖的方式实现防抖拍摄图像。可以理解的是，所述图像获取装置20的类型在此仅仅作为示例性的，而非限制。因此，其他类型的结构和安装方式也可应用于此。
如图12至图14所示，所述图像获取装置20包括一云台21和至少一摄像机22，其中所述云台21安装所述摄像机22至所述收割机主机10，所述云台21固定所述摄像机22的安装位置。所述云台21的底端被装载至所述收割机主机10，藉由所述收割机主机10固定所述云台21，其中所述云台21的上端被设置连接于所述摄像机22。所述摄像机22被所述云台21支撑而保持相对的平衡，以便稳定地拍摄图像或影像。所述摄像机22在所述云台21的支撑作用下拍摄所述收割机主机10周围的图像或影像，其中所述摄像机22基于所述云台21的安装位置为基准拍摄所述收割机主机10视野范围内的图像。可以理解的是，所述图像获取装置20的所述摄像机22基于所述收割机主机10的位置，通过拍照的方式获取至少一视觉图像。换言之，所述图像获取装置20的所述摄像机22是基于所述收割机主机10的视野范围内获取所述图像，从而避免摄像装置20的位置与收割机主机10位置变化，而导致的图像数据不准的问题。
值得一提的是，自动驾驶汽车，在自动驾驶模式下为识别获取车辆行走的路线，需要获取精确的车辆定位信息，通常需要高精度的卫星定位信息，并且所述 自动驾驶汽车需时刻更新道路中存在的障碍物信息、路面车辆信息、以及路面行人等信息，以在高速运行状态下实现自动驾驶功能。本发明的所述收割机获取的图像是对应于农田中的农作物谷物的图像数据信息，其中所述图像是基于车辆当前位置获取的所述车辆周边的图像。所述收割机不需要太高精度的卫星定位信息，只需要普通米级精度的卫星定位(GPS定位或者北斗定位等)。相应地，所述收割机所获取和处理的图像与自动驾驶汽车不同，因此，所述收割机所形成的路径规划和驾驶方式也不相同。可以理解的是，本发明的收割机基于视觉的识别所述农田的区域和自动驾驶功能与自动驾驶汽车的识别模式不同。
所述图像获取装置20的所述云台21进一步包括一云台固定件211和至少一云台移动件212，其中所述云台移动件212被可活动地连接至所述云台固定件211。所述云台固定件211被固定地设置于所述收割机主机10，其中所述摄像机22被安装至所述云台移动件212。所述云台21的所述云台移动件212活动地支撑所述摄像机22，以使所述摄像机22在所述收割机主机10抖动时保持相对位置的稳定，从而拍设置清晰的图像。
换言之，当所述收割机主机10在震动或抖动时，比如在农田中进行收割作业时产生的机械震动或抖动，所述云台21的所述云台固定件21与所述收割机主机10同步地震动，其中所述云台21的所述云台移动件212相对于所述云台固定件211运动，中和所述云台固定件211产生的震动，从而保持所述摄像机22的位置的稳定。详细地说，所述云台移动件212中和所述运动固定件211上下方向、左右方向、以及前后方向的抖动或震动，以保持所述摄像机22拍照位置的稳定，进而拍摄出稳定的图像信息。
如图14所示，所述图像获取装置20的所述摄像机22被设置于所述云台21的所述云台移动件212，其中所述摄像机22被固定地或可运动地安装至所述云台21的所述云台移动件212。优选地，所述摄像机22被可运动地设置于所述运动移动件212，其中所述摄像机22可基于所述云台移动件212的上端转动，以拍摄不同视野方向的图像。可选地，所述摄像机22被固定地安装至所述云台移动件212的上端，其中所述摄像机22在所述云台21的固定支撑作用下拍摄指定视野范围内的图像，比如拍摄所述收割机主机10前方视野内的图像。
所述摄像机22包括一摄像机主体221和至少一摄像机驱动装置222，其中所述摄像机驱动装置222驱动所述摄像机主体221的移动，以拍摄不同方向视野的 图像。所述摄像机主体221被可活动地设置于所述云台移动件212，其中所述摄像机主体221在所述摄像机驱动装置222的驱动作用下可在上下方向转动，以拍摄所述收割机主机10远方和附近位置处的农田和农作物的图像。可以理解的是，所述摄像机主体221被所述摄像机驱动装置222驱动向下转动时，所述摄像机主体221拍摄所述收割机主机10近处的图像，以便清晰识别所述图像中农作物的信息。当所述摄像机主体221被所述摄像机驱动装置222驱动而向上转动时，所述摄像机主体221拍摄所述收割机主机10远处的图像，以便通过所述图像识别出所述农田的作业区域和田边界区域。
所述摄像机驱动装置222驱动所述摄像机主体221在左右方向转动，以供所述摄像机主体221拍摄所述收割机主机10的左侧和右侧图像，以便识别所述农田的未作业区域100和已作业区域200，以及田边界区域300。
依照本发明说明书附图之图15所示，示出了所述图像获取装置20被安装于所述收割机主机10的几个可选的安装方式和安装位置。在本发明的第二较佳实施例中，所述收割机的所述图像获取装置20被设置于所述收割机主机10的前侧位置处、上部顶端处、左侧、右侧、以及后部位置处等。可以理解的是，所述图像获取装置20的安装位置不同，所拍摄的图像不同，从所述图像中识别出的信息也不相同。可以理解的是，被设置于所述收割机主机10前侧的所述图像获取装置20拍摄所述收割机主机10前方的图像，在所述收割机向前行驶作业时，所述收割机主机10前侧的所述图像获取装置20拍摄到所述收割机主机10的作业情况，以便根据拍摄的作业情况调整所述收割机主机10的行驶路径，作业参数等。
被设置于所述收割机主机10后侧的所述图像获取装置20拍摄所述收割机主机10后方的图像，在所述收割机向前行驶作业时，所述图像获取装置20拍摄的所述已作业区域200的图像。通过识别所述收割机主机10后侧的所述图像获取装置20拍摄的所述已作业区域200的图像，识别出所述收割机主机10的收割作业是否合格，以便于调整所述收割机主机10的作业参数。可以理解的是，通过设置于所述收割机主机10后侧的所述图像获取装置20拍摄的图像，所述收割机主机10识别出已作业区域200的农作物是否收割完全，是否遗留农作物颗粒等。所述收割机主机10被根据所述图像中识别出的信息调整作业参数，进而改善收割作业。值得一提的是，在倒车行驶时，所述图像获取装置20拍摄的图像为驾 驶人员提供倒车影像。
被设置于所述收割机主机10上部顶端的所述图像获取装置20拍摄所述收割机主机10远距离的图像，以便基于所述图像识别出农田的作业区域，田边界区域等。优选地，设置于所述收割机主机10上部顶端的所述图像获取装置20为可转动的云台摄像机。
相应地，被设置于所述收割机主机10左侧或右侧的所述图像获取装置20拍摄所述收割机主机10左侧或右侧的图像。基于所述收割机主机10左侧或右侧的图像，识别出所述图像中农田中农作物情况，以便识别所述未作业区域100、所述已作业区域200、以及所述田边界区域300。
如图12和图14所示，所述收割机进一步包括一图像处理***30、一定位装置40、以及一导航***50，其中所述图像处理***30、所述定位装置40、以及所述导航***50被设置于所述收割机主机10。所述定位装置40获取所述收割机主机10的位置信息，和将获取的位置信息传输至所述收割机主机10。所述导航***50基于所述定位装置40的定位信息为所述收割机主机10提供导航信息。所述图像处理***30基于所述图像获取装置20获取的所述农田的所述图像，从图像中识别出所述未作业区域100、所述已作业区域200、以及所述田边界区域300。
优选地，所述图像处理***30利用图像分割识别技术从图像中识别出所述未作业区域100、所述已作业区域200、以及所述田边界区域300。可以理解的是，所述图像处理***30还可以通过其他方式识别出所述图像中的区域和边界信息。因此，在本发明的第二较佳实施例中，所述图像处理***30识别图像的方式在此仅仅作为示例性的，而非限制。
如图16A和图16B所示，所述图像处理***30基于所述图像获取装置20拍摄的所述收割机主机10周围的图像，识别出所述图像中农田的区域、田边界，和识别出农田中农作物的种类，农作物的高度、颗粒饱满度、茎秆粗细大小等信息。
值得一提的是，所述图像处理***30选自基于阈值的分割方法、基于区域的分割方法、基于边缘的分割方法以及基于特定理论的分割方法等其中的任一分割识别方法对所述图像获取装置20获取的图像进行分割识别，以识别出所述图像中的区域和边界。优选地，所述图像处理***30利用深度学习算法对所述图 像分割识别和对所述图像进行区域划分和边界的限定。换言之，所述图像处理***30利用深度学习算法识别所述图像中对应的农田的区域和边界，以供所述收割机主机10根据识别划分的区域和边界行驶和进行作业。可更优选地，所述图像处理***30利用的深度学习算法为卷积神经网络算法的图像分割识别技术从图像中识别出对应农田中的所述未作业区域100、已作业区域200、以及所述田边界区域300。
值得一提地是，所述图像处理***30利用的处理算法在此仅仅作为示例性质的，而非限制。因此，所述图像处理***30还可利用其它算法对获取的图像进行分割识别，识别出图像中农田的区域和边界。
可以理解的是，所述图像处理***30是设置于所述收割机主机10的一图像处理器，其中所述图像处理器接收所述图像获取装置20拍摄的图像或影像，和识别出所述图像或影像中的信息。所述收割机主机10根据所述图像处理***30识别出的信息对应地操作控制行驶路径和调节作业的参数。
如图12和图14所示，所述收割机主机10进一步包括一车辆主体11，设置于所述车辆主体11的一作业***12，以及一驾驶控制***13，其中所述作业***12传动地连接于所述车辆主体11，其中所述车辆主体11带动所述作业***12工作，驱动所述作业***12进行收割农作物的作业。所述驾驶控制***13控制所述车辆主体11的行驶和控制所述作业***12的作业。值得一提的是，所述驾驶控制***13具有一无人驾驶模式和一操作驾驶模式。当所述收割机处于所述无人驾驶模式时，所述驾驶控制***13控制所述车辆主体11自动地运行和所述作业***12的作业。相应地，当收割机处于所述操作驾驶模式时，所述驾驶控制***允许驾驶人员通过人工操作的方式操作所述车辆主体11的运行和控制所述作业***的作业。
在本发明的第二较佳实施例中，所述驾驶控制***13控制所述车辆主体11的行驶和控制所述作业***12的收割作业。换言之，所述驾驶控制***13控制所述车辆主体11在行驶的过程中所述作业***12作业参数的调整。所述驾驶控制***13获取所述图像处理***30识别所述图像中的农作物的种类、农作物高度、颗粒饱满程度、农作物茎秆的直径大小等信息，和基于获取的所述信息调整所述作业***12的作业参数，比如，调整所述作业***12作业速度，作业的宽幅，作业的高度，调整后处理的参数等。
所述作业***12进一步包括至少一收割装置121，至少一输送装置122，以及至少一后处理装置123，其中所述输送装置122被设置能够接收所述收割装置121收割得到的作物，和将所述作物输送至所述后处理装置123，以供所述后处理装置123对所述作物进行后处理。所述作业***的所述收割装置121、所述输送装置122、以及所述后处理装置123分别被传动地设置连接于所述车辆主体11，藉由所述车辆主体11驱动所述作业***12的所述收割装置121、所述输送装置122、以及所述后处理装置123运行和作业。可以理解的是，所述后处理装置123被实施为作物的收割后的后续处理装置，例如，谷物收割机，所述后处理装置123为脱粒装置，割草设备中所述后处理装置123被实施为打包装置，当所述收割机为蔬菜水果的收获设备，所述后处理装置123被实施为蔬菜水果的筛选、存储装置。
所述驾驶控制***13根据所述图像处理***30识别出的所述图像信息控制所述收割装置121宽幅、收割高度、以及收割速度。可以理解的是，当农田中农作物的密度大时，所述图像获取装置20拍摄的所述农田中农作物的信息被所述图像处理***30识别，其中所述驾驶控制***13根据所述图像处理***30识别出的所述图像信息控制减小所述收割装置121的收割幅度、提升收割高度、以及减小收割速度等任一作业参数。
所述驾驶控制***13根据所述图像处理***30识别出的所述图像信息控制所述输送装置122的输送速度，输送功率等。可以理解的是，当农田中农作物的茎秆粗大，农作物的高度高，密度大时，所述图像获取装置20拍摄的所述农田中农作物的信息被所述图像处理***30识别，其中所述驾驶控制***13根据所述图像处理***30识别出的所述图像信息控制提升所述输送装置122的输送速度，提升输送功率等作业参数。
所述驾驶控制***13根据所述图像处理***30识别出的所述图像信息控制所述后处理装置123的后处理参数。可以理解的是，当农田中农作物的颗粒饱满程度、颗粒大小、水分含量、干湿程度、农作物果实的种类等。可以理解的是，所述图像处理***30识别出所述农田中所述农作物的农作物信息，其中所述驾驶控制***13根据所述图像处理***30识别出的所述图像信息调整所述后处理装置的后处理参数，比如吹风功率，后处理仓的转动速度等参数。
参照本发明说明书附图之图17所示，依照本发明第二较佳实施例的所述收 割机的一图像获取装置20A的另一可选实施方式在接下来的描述中被阐明。所述图像获取装置20A在本可选实施方式中是通过对摄像机内部控制镜头的视角和变焦，从而实现镜头拍照防止抖动。
相应地，所述图像获取装置20A包括一相机安装机构21A和至少一摄像机22A，其中所述相机安装机构21A将所述摄像机22A装载至所述收割机主机10。所述相机安装机构21A的底端被装载至所述收割机主机10，藉由所述收割机主机10固定所述相机安装机构21A，其中所述相机安装机构21A的上端被设置连接于所述摄像机22A。所述摄像机22A被所述相机安装机构21A支撑而保持相对的平衡，以便稳定地拍摄图像或影像。所述摄像机22A在所述相机安装机构21A的支撑作用下拍摄所述收割机主机10周围的图像或影像，其中所述摄像机22A基于所述相机安装机构21A的安装位置为基准拍摄所述收割机主机10视野范围内的图像。
可以理解的是，所述图像获取装置20A的所述摄像机22A基于所述收割机主机10的位置，通过拍照的方式获取至少一视觉图像。换言之，所述图像获取装置20A的所述摄像机22A是基于所述收割机主机10的视野范围内获取所述图像，从而避免摄像装置20A的位置与收割机主机10位置变化，而导致的图像数据不准的问题。
依照本发明的另一方面，本发明进一步提供一收割机的自动驾驶方法，其中所述自动驾驶方法包括如下方法步骤：
(a)获取至少一图像，和识别所述图像中农田的区域和田边界；
(b)基于所述识别信息，规划出至少一行驶规划路径603；以及
(c)控制所述收割机主机10按照所述行驶规划路径603自动地行驶。
上述自动驾驶方法步骤中，所述驾驶控制***13基于所述图像处理***30识别的区域信息和田边界控制所述收割机主机10的驾驶和作业。
上述自动驾驶方法的步骤(a)进一步包括：识别出所述图像中对应农田中农作物的信息，其中所述农作物的信息包括农作物种类，农作物的高度，颗粒饱满度等信息。
上述自动驾驶方法步骤(b)进一步包括步骤：
(b.1)识别划分出所述图像对应农田的区域和边界；以及
(b.2)基于识别的所述区域规划出至少一行驶规划路径603。
上述自动驾驶方法的步骤(b.1)进一步包括步骤：利用图像分割技术分割所述图像，和识别划分所述图像的区域。
在上述自动驾驶方法的步骤(a)中，基于所述收割机主机10的位置和行驶方向，实时地拍摄所述收割机主机10周围的图像信息。换言之，所述图像获取装置20实时地拍摄所述收割机主机10位置附近的图像。
在上述自动驾驶方法的步骤(b)中，所述图像处理***利用图像分割技术分割所述图像信息，和识别划分所述图像的区域为所述未作业区域100、所述已作业区域200、以及所述田边界区域300。相应地，所述自动驾驶方法的步骤(b.1)，进一步包括如下步骤：
分割所述图像为多个所述像元区域301，和归一化所述像元区域301的像素值为一数组；
提取每一数组对应的所述像元区域301的特征；以及
基于所述像元区域301对应的特征，输出所述图像的分类标签。
可以理解的是，所述分类标签对应于所述未作业区域100、所述已作业区域200、以及所述田边界区域300。
上述自动驾驶方法的步骤(b.2)进一步包括步骤：基于所述收割机主机10的定位信息，所述图像处理***30识别所述图像的区域规划信息，以及所述导航***50的导航信息，得出所述行驶规划路径603。
上述自动驾驶方法的步骤(b.2)进一步包括步骤：基于所述图像处理***30识别所述图像中的农作物的信息调整所述收割机主体10的行驶方向，以形成一车辆行驶路径604。
上述自动驾驶方法进一步包括：步骤(b.3)对比所述图像处理***30识别出的区域划分和区域边界范围与之前的区域边界范围是否保持一致，若不能保持一致，则调整所述图像对应的区域划分和区域边界范围，若能够保持一致，则保持区域划分和边界范围不变。
相应地，在上述方法步骤(c)中，所述驾驶控制***13根据所述收割机主机10的定位信息、所述图像处理***30得到的所述农田的区域规划信息、以及所述导航信息，控制所述收割机主机10的所述车辆主体11行驶。
在上述自动驾驶方法中，进一步包括步骤：(d)基于所述图像的识别信息，调整所述收割机主机10的作业***12的作业参数。
本领域的技术人员应理解，上述描述及附图中所示的本发明的实施例只作为举例而并不限制本发明。本发明的目的已经完整并有效地实现。本发明的功能及结构原理已在实施例中展示和说明，在没有背离所述原理下，本发明的实施方式可以有任何变形或修改。

Claims (25)

  1. 一收割机，其特征在于，包括：
    一收割机主机；
    至少一图像获取装置，其中所述图像获取装置被设置于所述收割机主机，所述图像获取装置拍摄所述收割机主机周围的图像，以及
    一图像处理***，其中所述图像处理***基于所述图像获取装置拍摄的图像识别出所述影像中的农田信息，其中所述收割机主机根据所述图像处理***识别出的所述农田信息自动地控制驾驶。
  2. 根据权利要求1所述的收割机，其中所述收割机进一步包括一路径规划***，其中所述路径规划***基于所述图像处理***识别的所述农田信息规划出至少一行驶规划路径，其中所述收割机主机根据所述路径规划***规划出的所述行驶规划路径控制驾驶。
  3. 根据权利要求1所述的收割机，其中所述图像处理***利用图像分割识别技术识别出所述图像中农田的信息，和基于识别出的信息规划所述图像中农田的区域。
  4. 根据权利要求3所述的收割机，其中所述图像处理***利用图像分割识别技术识别出所述图像中农作物信息，以供所述收割机主机基于识别出的信息自动地调整作业参数。
  5. 根据权利要求1所述的收割机，其中所述图像获取装置为防抖云台摄像装置，所述图像获取装置被装载于所述收割机主机，基于所述收割机主机的位置以拍照的方式拍摄所述收割机主机周围的图像。
  6. 根据权利要求4所述的收割机，其中所述图像获取装置为防抖云台摄像装置，所述图像获取装置被装载于所述收割机主机，基于所述收割机主机的位置以拍照的方式拍摄所述收割机主机周围的图像。
  7. 根据权利要求6所述的收割机，其中所述图像获取装置为机械防抖云台装置，所述图像获取装置包括一云台和至少一摄像机，其中所述云台将所述摄像机安装至所述收割机主机，所述摄像机被设置于所述云台，藉由所述云台支撑所述摄像机保持平衡。
  8. 根据权利要求6所述的收割机，其中所述图像获取装置为电子云台装置，所述图像获取装置通过控制镜头的视角和变焦，从而防止所述图像获取装置镜头拍照抖动。
  9. 根据权利要求7所述的收割机，其中所述图像获取装置被设置于所述收割机主机的前部、所述收割机主机的顶部、所述收割机主机的左侧、右侧、或所述收割机主机的后部。
  10. 根据权利要求8所述的收割机，其中所述图像获取装置被设置于所述收割机主机的前部、所述收割机主机的顶部、所述收割机主机的左侧、右侧、或所述收割机主机的后部。
  11. 根据权利要求3所述的收割机，其中所述图像处理***进一步包括：
    一图像分割模块，其中所述图像分割模块分割所述图像为多个像元区域，其中每一所述像元区域包括至少一像素单元；
    一特征化模块，其中所述特征化模块基于所述像元区域的所述像素单元提取每一像元区域对应的特征；以及
    一区域划分模块，其中所述区域规划模块根据所述像元区域的特征识别和划分所述图像的区域。
  12. 根据权利要求4所述的收割机，其中所述图像处理***进一步包括：
    一图像分割模块，其中所述图像分割模块分割所述图像为多个像元区域，其中每一所述像元区域包括至少一像素单元；
    一特征化模块，其中所述特征化模块基于所述像元区域的所述像素单元提取每一像元区域对应的特征；以及
    一区域划分模块，其中所述区域规划模块根据所述像元区域的特征识别和划分所述图像的区域。
  13. 根据权利要求3所述的收割机，其中所述收割机进一步包括一定位装置和一导航***，所述定位装置和所述导航***被设置于所述收割机主机，其中所述定位装置获取所述收割机主机的位置信息，其中所述导航***为所述谷物处理主体提供导航信息。
  14. 根据权利要求13所述的收割机，其中所述路径规划***进一步包括：
    一作业区域设置模块，其中所述作业区域设置模块设定所述农田的边界区域得到的所述农田的作业区域和所述作业边界；和
    一行驶路径规划模块，其中所述基于所述收割机主机的定位信息，所述图像处理***识别所述图像的区域规划信息，以及所述导航***的导航信息，得出至少一行驶规划路径。
  15. 根据权利要求1所述的收割机，其中所述收割机主机包括一车辆主体，设置于所述车辆主体的至少一作业***，以及一驾驶控制***，所述车辆主体驱动所述作业***运行，其中所述驾驶控制***控制所述车辆主体的运行和控制所述作业***的作业参数。
  16. 根据权利要求15所述的收割机，其中所述驾驶控制***获取所述图像处理***识别的所述图像获取装置拍摄的图像的信息，自动地控制所述车辆主体的行驶路线和控制所述作业***的作业参数，以实现无人自动驾驶和收割作业。
  17. 一收割机的自动驾驶方法，其特征在于，其中所述自动驾驶方法包括如下步骤：
    (a)获取至少一图像，和识别所述图像中农田的区域和田边界；
    (b)基于所述识别信息，规划出至少一行驶规划路径；以及
    (c)控制所述收割机主机按照所述行驶规划路径自动地行驶。
  18. 根据权利要求17所述的自动驾驶方法，其中上述自动驾驶方法的步骤
    (a)进一步包括：识别出所述图像中对应农田中农作物的信息，其中所述农作物的信息包括农作物种类，农作物的高度，颗粒饱满度等信息。
  19. 根据权利要求17所述的自动驾驶方法，其中上述自动驾驶方法步骤(b)进一步包括步骤：
    (b.1)识别划分出所述图像对应农田的区域和边界；以及
    (b.2)基于识别的所述区域规划出至少一行驶规划路径。
  20. 根据权利要求19所述的自动驾驶方法，其中上述自动驾驶方法的步骤(b.1)进一步包括步骤：利用图像分割技术分割所述图像，和识别划分所述图像的区域。
  21. 根据权利要求17所述的自动驾驶方法，其中在上述自动驾驶方法的步骤(b)中，所述图像处理***利用图像分割技术分割所述图像信息，和识别划分所述图像的区域为所述未作业区域、所述已作业区域、以及所述田边界区域。
  22. 根据权利要求20所述的自动驾驶方法，其中所述自动驾驶方法的步骤(b.1)进一步包括如下步骤：
    分割所述图像为多个所述像元区域，和归一化所述像元区域的像素值为一数组；
    提取每一数组对应的所述像元区域的特征；以及
    基于所述像元区域对应的特征，输出所述图像的分类标签。
  23. 根据权利要求22所述的自动驾驶方法，其中上述自动驾驶方法的步骤(b.2)进一步包括步骤：基于所述收割机主机的定位信息，所述图像的区域规划信息，以及所述导航***的导航信息，规划出所述行驶规划路径。
  24. 根据权利要求23所述的自动驾驶方法，其中上述自动驾驶方法进一步包括：步骤(b.3)对比所述图像处理***识别出的区域划分和区域边界范围与之前的区域边界范围是否保持一致，若不能保持一致，则调整所述图像对应的区域划分和区域边界范围，若能够保持一致，则保持区域划分和边界范围不变。
  25. 根据权利要求17所述的自动驾驶方法，其中上述自动驾驶方法进一步包括步骤：(d)基于所述图像的识别信息，调整所述收割机主机的作业***的作业参数。
PCT/CN2019/107551 2018-12-29 2019-09-24 收割机及其自动驾驶方法 WO2020134236A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP2021538493A JP2022516898A (ja) 2018-12-29 2019-09-24 ハーベスター及びその自動運転方法

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
CN201822267500.8 2018-12-29
CN201811638418.X 2018-12-29
CN201811638418.XA CN109588107A (zh) 2018-12-29 2018-12-29 收割机及其自动驾驶方法
CN201822267500.8U CN209983105U (zh) 2018-12-29 2018-12-29 收割机

Publications (1)

Publication Number Publication Date
WO2020134236A1 true WO2020134236A1 (zh) 2020-07-02

Family

ID=71128553

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/107551 WO2020134236A1 (zh) 2018-12-29 2019-09-24 收割机及其自动驾驶方法

Country Status (2)

Country Link
JP (1) JP2022516898A (zh)
WO (1) WO2020134236A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113311861A (zh) * 2021-05-14 2021-08-27 国家电投集团青海光伏产业创新中心有限公司 光伏组件隐裂特性的自动化检测方法及其***
EP3907469A4 (en) * 2019-01-04 2022-09-14 FJ Dynamics Technology Co., Ltd AUTOMATIC DRIVE SYSTEM FOR GRAIN PROCESSING, AND AUTOMATIC DRIVE METHOD AND PATH PLANNING METHOD THEREOF

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR102449993B1 (ko) * 2022-07-08 2022-10-06 주식회사 긴트 자율주행 농사용 차량의 작업 보조 방법 및 시스템

Citations (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5911669A (en) * 1996-04-19 1999-06-15 Carnegie Mellon University Vision-based crop line tracking for harvesters
CN101292572A (zh) * 2005-10-25 2008-10-22 迪尔公司 用于车辆导航的农作物属性图输入
EP2401906A1 (de) * 2010-07-01 2012-01-04 CLAAS Selbstfahrende Erntemaschinen GmbH Vorrichtung zur Erfassung und Bestimmung der Zusammensetzung von Schüttgut
CN103914071A (zh) * 2014-04-02 2014-07-09 中国农业大学 一种用于谷物联合收割机的视觉导航路径识别***
CN207045263U (zh) * 2016-08-10 2018-02-27 优势拓展（北京）科技有限公司 车载吸盘全景摄像机平台
CN207166612U (zh) * 2017-08-24 2018-03-30 深圳市维海德技术股份有限公司 基于一体化事件检测摄像机的智能交通***
CN108010033A (zh) * 2016-11-02 2018-05-08 哈尔滨派腾农业科技有限公司 一种农田场景图像采集和处理方法
CN108202667A (zh) * 2016-12-19 2018-06-26 株式会社久保田 作业车
CN108777938A (zh) * 2016-03-29 2018-11-09 洋马株式会社 联合收割机
CN109588107A (zh) * 2018-12-29 2019-04-09 丰疆智慧农业股份有限公司 收割机及其自动驾驶方法
CN109631903A (zh) * 2019-01-04 2019-04-16 丰疆智慧农业股份有限公司 谷物处理自动驾驶***及其自动驾驶方法和路径规划方法
CN109716917A (zh) * 2019-01-04 2019-05-07 丰疆智慧农业股份有限公司 带有云台摄像装置的收割机
CN110209154A (zh) * 2019-04-09 2019-09-06 丰疆智能科技股份有限公司 自动收割机的残留收割路径规划***及其方法
CN110209156A (zh) * 2019-04-09 2019-09-06 丰疆智能科技股份有限公司 自动收割机的行驶路径规划***及其方法

Family Cites Families (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH01211408A (ja) * 1988-02-18 1989-08-24 Yanmar Agricult Equip Co Ltd 農作業機における作物列検出装置
JP2795454B2 (ja) * 1989-03-16 1998-09-10 ヤンマー農機株式会社 カラー情報による境界線の検出方法
JP2740268B2 (ja) * 1989-06-16 1998-04-15 ヤンマー農機株式会社 走行作業機の操舵制御における画像情報の２値化処理装置
JP4340367B2 (ja) * 1999-01-27 2009-10-07 株式会社リコー 画像分類装置およびその装置としてコンピュータを機能させるためのプログラムを記録したコンピュータ読み取り可能な記録媒体
JP3877301B2 (ja) * 2002-06-11 2007-02-07 ヤンマー農機株式会社 農業用作業車
JP2006121952A (ja) * 2004-10-27 2006-05-18 Iseki & Co Ltd コンバイン
JP4624884B2 (ja) * 2005-08-08 2011-02-02 株式会社クボタ 作業車の画像処理装置
JP5818448B2 (ja) * 2011-02-02 2015-11-18 ヤンマー株式会社 苗植付機
JP6635909B2 (ja) * 2016-12-28 2020-01-29 本田技研工業株式会社 作業機、制御装置及び制御用プログラム
JP2018186728A (ja) * 2017-04-28 2018-11-29 井関農機株式会社 コンバイン

Patent Citations (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5911669A (en) * 1996-04-19 1999-06-15 Carnegie Mellon University Vision-based crop line tracking for harvesters
CN101292572A (zh) * 2005-10-25 2008-10-22 迪尔公司 用于车辆导航的农作物属性图输入
EP2401906A1 (de) * 2010-07-01 2012-01-04 CLAAS Selbstfahrende Erntemaschinen GmbH Vorrichtung zur Erfassung und Bestimmung der Zusammensetzung von Schüttgut
CN103914071A (zh) * 2014-04-02 2014-07-09 中国农业大学 一种用于谷物联合收割机的视觉导航路径识别***
CN108777938A (zh) * 2016-03-29 2018-11-09 洋马株式会社 联合收割机
CN207045263U (zh) * 2016-08-10 2018-02-27 优势拓展（北京）科技有限公司 车载吸盘全景摄像机平台
CN108010033A (zh) * 2016-11-02 2018-05-08 哈尔滨派腾农业科技有限公司 一种农田场景图像采集和处理方法
CN108202667A (zh) * 2016-12-19 2018-06-26 株式会社久保田 作业车
CN207166612U (zh) * 2017-08-24 2018-03-30 深圳市维海德技术股份有限公司 基于一体化事件检测摄像机的智能交通***
CN109588107A (zh) * 2018-12-29 2019-04-09 丰疆智慧农业股份有限公司 收割机及其自动驾驶方法
CN109631903A (zh) * 2019-01-04 2019-04-16 丰疆智慧农业股份有限公司 谷物处理自动驾驶***及其自动驾驶方法和路径规划方法
CN109716917A (zh) * 2019-01-04 2019-05-07 丰疆智慧农业股份有限公司 带有云台摄像装置的收割机
CN110209154A (zh) * 2019-04-09 2019-09-06 丰疆智能科技股份有限公司 自动收割机的残留收割路径规划***及其方法
CN110209156A (zh) * 2019-04-09 2019-09-06 丰疆智能科技股份有限公司 自动收割机的行驶路径规划***及其方法

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3907469A4 (en) * 2019-01-04 2022-09-14 FJ Dynamics Technology Co., Ltd AUTOMATIC DRIVE SYSTEM FOR GRAIN PROCESSING, AND AUTOMATIC DRIVE METHOD AND PATH PLANNING METHOD THEREOF
CN113311861A (zh) * 2021-05-14 2021-08-27 国家电投集团青海光伏产业创新中心有限公司 光伏组件隐裂特性的自动化检测方法及其***

Also Published As

Publication number Publication date
JP2022516898A (ja) 2022-03-03

Similar Documents

Publication Publication Date Title
WO2020140491A1 (zh) 谷物处理自动驾驶***及其自动驾驶方法和路径规划方法
US11074680B2 (en) Management and display of object-collection data
KR102618797B1 (ko) 콤바인, 포장 영농 맵 생성 방법, 포장 영농 맵 생성 프로그램 및 포장 영농 맵 생성 프로그램이 기록된 기록 매체
RU2747303C2 (ru) Система для управления рабочим параметром уборочной жатки
CN109588107A (zh) 收割机及其自动驾驶方法
WO2020134236A1 (zh) 收割机及其自动驾驶方法
US11319067B2 (en) Drone for capturing images of field crops
CN209983105U (zh) 收割机
WO2020140492A1 (zh) 谷物处理自动驾驶***、自动驾驶方法以及自动识别方法
CN109716917A (zh) 带有云台摄像装置的收割机
WO2020140490A1 (zh) 带有云台摄像装置的收割机
CN210130123U (zh) 带有云台摄像装置的收割机
WO2022123889A1 (ja) 作業車、作物状態検出システム、作物状態検出方法、作物状態検出プログラム、及び作物状態検出プログラムが記録されている記録媒体
JP7397880B2 (ja) 農業刈り作業機械のための画像収集装置及びその処理方法
WO2020262287A1 (ja) 農作業機、自動走行システム、プログラム、プログラムを記録した記録媒体、及び方法
CN210466135U (zh) 农割作业机器
RU2774651C1 (ru) Система автоматического вождения для переработки зерна, способ автоматического вождения и способ планирования траектории
US20230230202A1 (en) Agricultural mapping and related systems and methods
WO2022071373A1 (ja) 収穫機
US20210019903A1 (en) System and method for determining an attribute of a plant
JP2023040743A (ja) 収穫機

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19906492

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2021538493

Country of ref document: JP

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 19906492

Country of ref document: EP

Kind code of ref document: A1