CN112784799A - AGV (automatic guided vehicle) backward pallet and obstacle identification method and device and AGV - Google Patents
AGV (automatic guided vehicle) backward pallet and obstacle identification method and device and AGV Download PDFInfo
- Publication number
- CN112784799A CN112784799A CN202110139701.3A CN202110139701A CN112784799A CN 112784799 A CN112784799 A CN 112784799A CN 202110139701 A CN202110139701 A CN 202110139701A CN 112784799 A CN112784799 A CN 112784799A
- Authority
- CN
- China
- Prior art keywords
- point cloud
- cloud data
- pallet
- agv
- obstacle
- Prior art date
- Legal status (The legal status 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 status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 76
- 230000008447 perception Effects 0.000 claims abstract description 35
- 238000012545 processing Methods 0.000 claims abstract description 34
- 230000004888 barrier function Effects 0.000 claims abstract description 13
- 238000001914 filtration Methods 0.000 claims description 35
- 239000013598 vector Substances 0.000 claims description 12
- 238000004590 computer program Methods 0.000 claims description 10
- 238000003860 storage Methods 0.000 claims description 9
- 238000000638 solvent extraction Methods 0.000 claims description 6
- 238000004519 manufacturing process Methods 0.000 abstract description 2
- 230000011218 segmentation Effects 0.000 description 15
- 238000005070 sampling Methods 0.000 description 10
- 238000010586 diagram Methods 0.000 description 8
- 230000008569 process Effects 0.000 description 8
- 230000006870 function Effects 0.000 description 6
- 238000004891 communication Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 5
- 238000011161 development Methods 0.000 description 3
- 230000018109 developmental process Effects 0.000 description 3
- 238000009826 distribution Methods 0.000 description 3
- 238000005259 measurement Methods 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 238000003708 edge detection Methods 0.000 description 2
- 230000005611 electricity Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000000513 principal component analysis Methods 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 1
- 238000005520 cutting process Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000005484 gravity Effects 0.000 description 1
- 238000003780 insertion Methods 0.000 description 1
- 230000037431 insertion Effects 0.000 description 1
- 238000003064 k means clustering Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000010845 search algorithm Methods 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012800 visualization Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Probability & Statistics with Applications (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a method and a device for identifying a backward pallet and a barrier of an AGV (automatic guided vehicle), and the AGV, wherein the method comprises the following steps: acquiring point cloud data behind the AGV, which is acquired by a perception sensor; processing the point cloud data, and determining pallet point cloud data and obstacle point cloud data in the point cloud data; based on the pallet point cloud data is discerned to the pallet after AGV, and based on obstacle point cloud data is discerned to the obstacle after AGV. According to the method, the pallet and the barrier can be identified simultaneously by only using one perception sensor, and the production cost of the AGV is reduced.
Description
Technical Field
The invention relates to the technical field of automatic control, in particular to an AGV and a method and a device for identifying a backward pallet and a barrier.
Background
The development and progress of science and technology has promoted the automation degree of various industries to be improved continuously. Especially in the fields of automatic storage and logistics, the automation control technology is paid more attention to and applied. An Automatic Guided Vehicle (AGV) is an important technology for Automatic handling and loading, and the application range and the technical level thereof have been rapidly developed.
When the AGV runs backwards, the AGV has the requirement of pallet identification because the automatic loading and unloading function of the pallet needs to be completed; meanwhile, when the AGV runs backwards, a barrier may be generated to block the AGV from running, so that the AGV also has the requirement of sensing the barrier.
Currently, identification of pallets and sensing of obstacles are accomplished by two sensors, respectively, which increases the density of sensors on the AGV body and the cost of the AGV.
Disclosure of Invention
The invention provides an AGV backward pallet and obstacle identification method and device and an AGV, which are used for solving the defect that two perception sensors are needed for AGV backward pallet and obstacle identification.
The invention provides an AGV backward pallet and obstacle identification method, which comprises the following steps:
acquiring point cloud data behind the AGV, which is acquired by a perception sensor;
processing the point cloud data, and determining pallet point cloud data and obstacle point cloud data in the point cloud data;
based on the pallet point cloud data is discerned to the pallet after AGV, and based on obstacle point cloud data is discerned to the obstacle after AGV.
According to the AGV backward pallet and obstacle identification method provided by the invention, the point cloud data are processed to determine pallet point cloud data and obstacle point cloud data in the point cloud data, and the method specifically comprises the following steps:
filtering ground point cloud data in the point cloud data to obtain alternative point cloud data;
and partitioning the alternative point cloud data to obtain the pallet point cloud data and the obstacle point cloud data.
According to the AGV backward pallet and obstacle identification method provided by the invention, ground point cloud data in the point cloud data are filtered out to obtain alternative point cloud data, and the method specifically comprises the following steps:
performing direct filtering on the point cloud data to determine point cloud data of the region of interest;
and filtering ground point cloud data in the point cloud data of the region of interest to obtain the alternative point cloud data.
According to the AGV backward pallet and obstacle identification method provided by the invention, the alternative point cloud data are segmented to obtain the pallet point cloud data and the obstacle point cloud data, and the method specifically comprises the following steps:
dividing the pallet point cloud data from the alternative point cloud data based on at least one of pallet color characteristics, pallet target object characteristics and pallet geometric characteristics, and obtaining residual point cloud data in the alternative point cloud data except the pallet point cloud data;
and determining obstacle point cloud data in the residual point cloud data based on a clustering algorithm.
According to the AGV backward pallet and obstacle identification method provided by the invention, the recognition of the AGV backward pallet based on the pallet point cloud data specifically comprises the following steps:
and processing the pallet point cloud data based on a normal vector estimation algorithm, a region generation algorithm or a point cloud clustering algorithm, and identifying the pose information of the AGV to the pallet.
According to the AGV backward pallet and obstacle identification method provided by the invention, the identification of the AGV backward obstacle based on the obstacle point cloud data specifically comprises the following steps:
and processing the obstacle point cloud data based on an obstacle perception algorithm or an Euclidean clustering algorithm, and identifying the state information of the obstacle behind the AGV.
The invention also provides an AGV backward pallet and obstacle recognition device, which comprises:
the acquisition module is used for acquiring point cloud data behind the AGV, which is acquired by the perception sensor;
the processing module is used for processing the point cloud data and determining pallet point cloud data and obstacle point cloud data in the point cloud data;
and the recognition module is used for recognizing the pallet after the AGV is processed by the pallet point cloud data and recognizing the obstacle after the AGV is processed by the obstacle point cloud data.
The present invention also provides an AGV, comprising: perception sensor and as above AGV after to pallet and barrier recognition device, the perception sensor with AGV is to pallet and barrier recognition device electricity connection.
The invention also provides electronic equipment, which comprises a memory, a processor and a computer program stored on the memory and capable of running on the processor, wherein the processor executes the program to realize the steps of any one of the methods for identifying the backward pallet and the obstacle of the AGV.
The present invention also provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of the AGV backward pallet and obstacle identification method as described in any of the above.
According to the method and the device for identifying the backward pallet and the obstacle by the AGV, and the AGV, the point cloud data is processed by acquiring the point cloud data acquired by the perception sensor, the pallet point cloud data and the obstacle point cloud data are segmented, the pallet point cloud data and the obstacle point cloud data are respectively identified, the pallet position information and the obstacle state information are obtained, and the function of identifying the backward pallet and the obstacle by the AGV by using only one perception sensor is realized. Moreover, the pallets and the obstacles can be identified simultaneously after the AGV, and the safety of the AGV in running after the AGV is started can be improved.
Drawings
In order to more clearly illustrate the technical solutions of the present invention or the prior art, the drawings needed for the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and those skilled in the art can also obtain other drawings according to the drawings without creative efforts.
FIG. 1 is a schematic flow chart illustrating a method for identifying a pallet and an obstacle backward from an AGV according to an embodiment of the present invention;
FIG. 2 is a schematic diagram illustrating a complete flow of a method for identifying a pallet and an obstacle backward from an AGV according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of an AGV backward pallet and obstacle recognition device according to the present invention;
FIG. 4 is a schematic diagram of an AGV according to the present invention;
fig. 5 is a schematic structural diagram of an electronic device provided in the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is obvious that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Currently, a camera or a multiline lidar is generally used for pallet identification and positioning, and a photoelectric sensor, an ultrasonic sensor or a single line lidar is used for obstacle sensing. Since the identification of the pallet and the sensing of the obstacle are done by different sensors, respectively, this increases the density of sensors on the AGV body and the cost of the AGV. Therefore, the embodiment of the invention provides an AGV backward pallet and obstacle identification method.
Fig. 1 is a schematic flowchart illustrating a method for identifying a pallet and an obstacle after AGV according to an embodiment of the present invention. As shown in fig. 1, the method includes:
s1, acquiring point cloud data acquired by a perception sensor and backward to the AGV;
s2, processing the point cloud data, and determining pallet point cloud data and obstacle point cloud data in the point cloud data;
s3, based on the pallet point cloud data is discerned to the pallet after AGV, and based on obstacle point cloud data is discerned to the obstacle after AGV.
Specifically, according to the method for identifying the backward pallet and the obstacle by the AGV provided in the embodiment of the present invention, the execution main body may be an AGV main controller, or may be a server independent from the AGV, for example, the execution main body may be a local server, a cloud server, or the like, and the local server may be a computer, a tablet computer, a smart phone, or the like, which is not specifically limited in the embodiment of the present invention.
Step S1 is performed first. Wherein, the perception sensor can be a laser radar. The laser radar is a scanning sensor adopting a non-contact laser ranging technology, and point cloud data is formed by collecting reflected light beams. The laser radar comprises a rotary laser radar, a plurality of laser beams are vertically arranged and rotate 360 degrees around an axis, each laser beam scans a plane, and a three-dimensional figure is displayed after longitudinal superposition. The laser radar also comprises a solid-state laser radar which adopts a phased array principle and consists of a plurality of fixed small light beams, and the phase and the amplitude of the light beams are generated through each array element point, so that the intensity of the light beams in the specified direction is enhanced, and the intensities in other directions are suppressed, and the change of the light beam direction is realized.
In another embodiment, the perception sensor may be a depth camera. The depth camera is composed of a passive camera and an active depth sensor. Unlike conventional cameras, depth cameras provide per-pixel depth information in addition to a color image (RGB), combining the advantages of active and passive sensors. Depth cameras calculate depth by emitting a light signal in a scene and analyzing the reflected light, but incident wave modulation is performed in space rather than time. The main function of the depth camera can be divided into two parts, namely the acquisition of RGB images and depth images. An RGB camera acquires an RGB image by measuring light and density, a depth image means that each pixel value in the image is the actual distance of the sensor from the object, and can be acquired by the depth sensor.
It should be noted that, the selection of the sensing sensor may be set as required, and the embodiment of the present invention is not particularly limited thereto, and as long as the sensor having the function of outputting the 3D point cloud is applicable to the embodiment of the present invention.
In the embodiment of the invention, the point cloud data acquired by the perception sensor is a basic three-dimensional model and refers to a set of vectors in a three-dimensional coordinate system. These vectors are usually expressed in terms of X, Y, Z three-dimensional coordinates and are generally used primarily to represent the shape of the external surface of an object. The point cloud data may represent RGB color, gray value, depth, segmentation result, and the like of one point in addition to the geometric position information represented by (X, Y, Z). The point cloud data is often obtained directly by measurement, each point corresponds to a measurement point, and other processing means are not used, so that the point cloud data contains the maximum information amount.
In the embodiment of the present invention, the point cloud data collected by the sensing sensor may be point cloud data based on a Robot Operating System (ROS). ROS is a widely used platform in the field of robotics research, and its development employs the open Berkeley Software Distribution (BSD) protocol. The ROS acts as a framework for programming the robot, linking the user's operating system with the ROS application developed by the user. The ROS adopts a distributed framework, so that modular modification and customization are facilitated, and the fault-tolerant capability of the system is improved. Meanwhile, the ROS supports multiple programming languages and has a huge open source community, so that the development of developers is facilitated.
Since the backward space of the AGV may include the pallet, the obstacle, the ground, and the like, the point cloud data collected by the sensing sensor may include at least one of the ground point cloud data, the pallet point cloud data, and the obstacle point cloud data backward from the AGV.
Then, step S2 is executed. Processing the point cloud data may include image enhancement, filtering, key point/edge detection, and other basic operations. After the basic operation is completed, the point cloud data can be divided, and the pallet point cloud data and the obstacle point cloud data in the point cloud data are determined through division.
In the embodiment of the invention, if the pallet point cloud data is not segmented, the situation that no pallet exists or the pallet is shielded after the AGV is shown, the point cloud data acquired by the perception sensor does not contain the pallet point cloud data, so that the pallet point cloud data cannot be segmented, and the point cloud data acquired by the perception sensor can be directly judged as the obstacle. Moreover, in order to avoid damage when the AGV touches an obstacle, the AGV may stop continuing to travel backwards.
Finally, step S3 is performed. Wherein, based on the pallet point cloud data that obtains of cutting apart, can discern the pallet to AGV back, can be including calculating the direction and the position etc. of pallet to the discernment of pallet. Meanwhile, the obstacle may be identified based on the obtained obstacle point cloud data, and the identification of the obstacle may include calculating the position, number, movement, and the like of the obstacle, which is not specifically limited in the embodiment of the present invention.
According to the AGV backward pallet and obstacle identification method provided by the embodiment of the invention, the point cloud data acquired by the perception sensor is acquired and processed, and the pallet point cloud data and the obstacle point cloud data can be determined by dividing the point cloud data. And respectively identifying the pallet point cloud data and the obstacle point cloud data according to the pallet point cloud data and the obstacle point cloud data. According to the method, the identification of the pallet and the obstacle can be completed simultaneously only by one sensing sensor, the density of the sensors on the AGV body is reduced, and the cost of the AGV is also reduced. Meanwhile, due to the fact that the point cloud data are accurate in perception of the 3D space, the recognition of the pallet and the obstacles behind the AGV can be more accurate through the point cloud data. Moreover, because can discern simultaneously to pallet and barrier behind the AGV, can avoid the AGV to have the barrier nevertheless to fail in time discernment and the potential safety hazard that exists behind the AGV, improve the security of AGV to the operation.
On the basis of the above embodiment, the method for identifying a pallet and an obstacle backward for an AGV according to the embodiment of the present invention processes the point cloud data to determine pallet point cloud data and obstacle point cloud data in the point cloud data, and specifically includes:
filtering ground point cloud data in the point cloud data to obtain alternative point cloud data;
and partitioning the alternative point cloud data to obtain the pallet point cloud data and the obstacle point cloud data.
Specifically, in the embodiment of the present invention, a Point Cloud Library (PCL) may be used to filter out ground Point Cloud data in the Point Cloud data. The PCL is a large cross-platform open source programming library used for processing two-dimensional/three-dimensional images and point cloud data. PCL realizes a large number of general algorithms and efficient data structures related to point cloud, and relates to point cloud acquisition, filtering, segmentation, feature extraction, visualization and the like. The PCL supports a variety of operating system platforms, such as Windows, Linux, Android, and the like.
And processing the acquired point cloud data acquired by the perception sensor by using a filter packaged in the PCL, and filtering ground point cloud data to obtain the alternative point cloud data.
Wherein the filter may be a pass-through filter. The straight-through filter is suitable for point cloud data with certain spatial characteristics distributed in space, such as point cloud data acquired by a linear optical structure scanning mode, the point cloud data is widely distributed along the z direction, but the distribution of the x and y directions is in a limited range, and at the moment, the range of the point cloud in the x or y direction can be determined by the straight-through filter, so that the function of quickly removing outliers is achieved.
In another embodiment, the filter may be a conditional filter. The conditional filter filters the point cloud data by setting filtering conditions, the point cloud data is left when the point cloud data is in a certain range, and the point cloud data is rejected when the point cloud data is not in the set condition range, so that the filtering effect is achieved.
After the ground point cloud data is filtered, alternative point cloud data can be obtained. At this time, the candidate point cloud data may include pallet point cloud data and/or obstacle point cloud data, and may also include noise point cloud data. If the alternative point cloud data at least comprises pallet point cloud data and obstacle point cloud data at the same time, the pallet point cloud data and the obstacle point cloud data can be segmented by utilizing a point cloud segmentation technology.
The point cloud data segmentation aims to extract point cloud data corresponding to different objects in the point cloud data. In the embodiment of the invention, the pallet point cloud data and the obstacle point cloud data in the alternative point cloud data are extracted, so that the purpose of independently processing the pallet point cloud data and the obstacle point cloud data can be realized.
The point cloud data segmentation can be realized by means of proximity search. The point cloud data is segmented by comparing certain characteristics of a certain point and the nearest point. Determining the neighborhood relationship between a point cloud data may be accomplished by a k-dimensional Tree (kd-Tree). And establishing a kd-Tree and searching in the kd-Tree. The kd-Tree establishing process is a continuous dividing process, firstly, the most sparse dimension is selected, then, the middle point on the dimension is found, and the first division is carried out perpendicular to the dimension. At this time, the k-dimensional hyperplane is divided into two, the sparsest dimension is found in the two sub-planes, and so on until the last point is also divided, and at this time, a continuous bisection tree is formed. Generally, the adjacent points of one point only need to search in the parent node and the child node, thereby greatly reducing the search scale of the adjacent points. And the kd-Tree can effectively judge the position of the nearest point of the insertion point. The kd-Tree algorithm is very important for low-level vision. In many cases, it is necessary to give a certain point and then look up the number of k-adjacent points or points within a certain radius.
According to the AGV backward pallet and obstacle identification method provided by the embodiment of the invention, the ground point cloud data is filtered, so that the interference of the ground point cloud data on the subsequent segmentation process is avoided, and the subsequent segmentation process is more efficiently and accurately realized. By dividing the alternative point cloud data, the pallet point cloud data and the obstacle point cloud data can be determined, the pallet point cloud data and the obstacle point cloud data can be conveniently and subsequently and independently processed, and the difficulty of subsequently identifying the pallet and the obstacle is reduced.
On the basis of the above embodiment, the method for identifying an AGV backward pallet and an obstacle according to the embodiment of the present invention filters ground point cloud data in the point cloud data to obtain alternative point cloud data, and specifically includes:
performing direct filtering on the point cloud data to determine point cloud data of the region of interest;
and filtering ground point cloud data in the point cloud data of the region of interest to obtain the alternative point cloud data.
Specifically, in the embodiment of the present invention, the point cloud data may be subjected to straight-through filtering in three directions of x, y, and z, and the dimensions x, y, and z and the value range under the dimensions are sequentially specified; secondly, traversing each point in the point cloud data, judging whether the value of the point on the specified dimension is in a specified value domain, and deleting the points of which the values are not in the value domain; and finally, finishing the traversal, and forming the point cloud data after the straight-through filtering by using the remained points. After the straight-through filtering, noise point cloud data can be filtered out, and point cloud data of the region of interest can be obtained.
In the embodiment of the invention, after the straight-through filtering processing, the outlier filtering can be carried out. Outlier filtering refers to performing statistical analysis on the field of each point in the point cloud data, and calculating the average distance from the point to all the nearby points. Assuming that the result is a gaussian distribution whose shape is determined by the mean and standard deviation, points whose mean distance is outside the standard range can be defined as outliers and deleted from the point cloud data. The standard range may be determined by the global distance average and the variance, or may be given, which is not specifically limited in the embodiment of the present invention.
In the embodiment of the invention, after the outlier filtering is carried out, the down-sampling can be carried out. Downsampling of point cloud data can be achieved by using a VoxelGrid class in a PCL library. Down-sampling is a technique of multi-rate digital signal processing or a process of reducing the signal sampling rate, and is generally used to reduce the data transmission rate or data size. The basic principle of the VoxelGrid class is to establish a three-dimensional voxel grid, grid-divide point clouds, the gravity centers of points falling in each small cubic block area represent all points in the grid, and the number of points for down-sampling can be controlled by controlling the side length of the grid.
In the embodiment of the invention, ground point cloud data in the point cloud data of the region of interest can be filtered out by using a random sampling consistency algorithm in PCL. The random sampling consistency algorithm may iteratively estimate the parameters of the mathematical model from a set of observed data sets containing "outliers". It is an uncertain algorithm with a certain probability to obtain a reasonable result. To increase the probability of a reasonable result, the number of iterations must be increased.
The random sampling consistency algorithm achieves this goal by iteratively selecting a set of random subsets in the data. The selected subset is assumed to be an intra-office point and verified by the steps of adapting a model to the assumed intra-office point, i.e. all unknown parameters can be calculated from the assumed intra-office point; all other data are tested using the model described above, and if a point is suitable for the estimated model, it is considered to be an in-office point; if enough points are classified as the assumed intra-office points, the estimated model is reasonable enough; then, all the assumed intra-office points are used to re-estimate the model, since it is estimated only by the initial assumed intra-office points; finally, the model is evaluated by estimating the error rate of the local interior point and the model. Through continuous iteration, the best model with the error rate within the allowable range can be obtained, all local interior points can be tested by using the model, all local exterior points are eliminated, and the filtering effect is achieved.
According to the AGV backward pallet and obstacle identification method provided by the embodiment of the invention, the region of interest can be determined through direct filtering, the range of point cloud data to be processed is reduced, and the subsequent operation is facilitated. The outlier filtering can further eliminate point cloud data outside the region of interest, and the subsequent filtering effect is improved. The time and space complexity of subsequent processing can be reduced by down-sampling, and the efficiency of the subsequent processing is improved. Ground point cloud data are filtered by using a random sampling consistency algorithm to obtain alternative point cloud data, and the subsequent point cloud data segmentation process is simplified.
On the basis of the above embodiment, the method for identifying an AGV backward pallet and a barrier according to the embodiment of the present invention is a method for segmenting the candidate point cloud data to obtain the pallet point cloud data and the barrier point cloud data, and specifically includes:
dividing the pallet point cloud data from the alternative point cloud data based on at least one of pallet color characteristics, pallet target object characteristics and pallet geometric characteristics, and obtaining residual point cloud data in the alternative point cloud data except the pallet point cloud data;
and determining obstacle point cloud data in the residual point cloud data based on a clustering algorithm.
In particular, in embodiments of the present invention, pallets may be marked in advance. For example, a specific color, such as red or green, may be marked in the middle of the pallet, or an object, such as a two-dimensional code, may be marked in the middle of the pallet. Meanwhile, the pallet has geometrical characteristics, so that the pallet point cloud data can be firstly segmented by utilizing at least one characteristic of color characteristics, pallet target object characteristics and pallet geometrical characteristics, residual point cloud data except the pallet point cloud data is obtained, and the pallet point cloud data is not contained in the residual point cloud data.
The remaining point cloud data may then be clustered based on a clustering algorithm to determine obstacle point cloud data that may be present in the remaining point cloud data. Clustering refers to dividing a data set into identical classes or clusters according to a certain criterion (e.g., distance) such that the similarity of data objects in a cluster is as large as possible, and the difference of data objects not in a cluster is also as large as possible. After clustering, the data of the same class are gathered together as much as possible, and different data are separated.
The clustering algorithm adopted in the embodiment of the invention can be K-means (K-means) clustering. The K-means clustering firstly randomly selects K objects, wherein K can be set according to needs, the invention does not limit the objects, and the respective central points of the objects are randomly initialized. The center point is the same length position as each data point vector. The distance of each data point to the center point is then calculated, and the class to which the data point is closest to which center point. And then, the average value of each class is recalculated and updated to be a new class center. The above steps are repeated until the center of each class does not change much after each iteration. It is also possible to randomly initialize the center point multiple times and then select the one that has the best run result.
In another embodiment, the clustering algorithm may also be bottom-up hierarchical clustering. The hierarchical clustering regards each datum as a single class, and the minimum distance between every two data is calculated; merging the two classes with the minimum distance into a new class; recalculating the distances between the new class and all classes; and repeating the steps of calculating the minimum distance between every two classes, combining the two classes with the minimum distance into a new class, and calculating the distances between the new class and all the classes until all the classes are finally combined into one class.
It should be noted that, when the alternative point cloud data is segmented, in addition to the pallet point cloud data segmented first, in the embodiment of the present invention, the obstacle point cloud data may be segmented first according to the movement state of the obstacle under the condition that the obstacle moves, and this is not specifically limited in the embodiment of the present invention.
According to the AGV backward pallet and obstacle identification method provided by the embodiment of the invention, the pallet point cloud data can be firstly divided based on at least one of the pallet color, the target characteristic or the self geometric characteristic, and then the obstacle point cloud is determined for the rest point cloud data based on a clustering algorithm. The method and the device achieve the purpose that only one sensor is needed to collect point cloud data, and the pallet point cloud data and the obstacle point cloud data can be obtained at the same time.
On the basis of the above embodiment, the method for identifying the AGV backward pallet and the obstacle provided in the embodiment of the present invention identifies the AGV backward pallet based on the pallet point cloud data, and specifically includes:
and processing the pallet point cloud data based on a normal vector estimation algorithm, a region generation algorithm or a point cloud clustering algorithm, and identifying the pose information of the AGV to the pallet.
Specifically, in the embodiment of the present invention, the normal vector estimation algorithm may select each point in the point cloud data, search the neighboring points of the point cloud data, fit the neighboring points including the point into a curved surface, perform Principal Component Analysis (PCA) on the points in the curved surface, and search for a feature vector corresponding to the minimum feature value, where the feature vector is a normal vector of the fitted curved surface. This fitting surface is the pallet surface, and then can determine the position appearance information of pallet after the AGV, for example including position and gesture.
In the embodiment of the present invention, the region generation algorithm may specifically use a region growing (region growing) class to implement region growing segmentation, and an output of the algorithm is a cluster set. The flow of the algorithm may be: setting an empty seed point sequence, an empty cluster array- > adding the initial seed point into the seed sequence- > searching the neighborhood point- > adding the current region- > if the included angle between the normal line of the field point and the normal line of the current seed point is smaller than the threshold value, adding the seed point sequence- > deleting the current seed point, and growing by using the new seed point until the seed sequence point is emptied.
In the embodiment of the present invention, the point cloud Clustering algorithm may use a Density-Based Clustering algorithm (DBSCAN). The DBSCAN algorithm assumes that the clustering structure can be determined by the closeness degree of sample distribution, examines the continuity between samples from the perspective of sample density, and continuously expands clustering clusters based on connectable samples to obtain a final clustering result. The DBSCAN algorithm defines clusters as the points where the density is the greatest, and divides areas with sufficient density into clusters, which can find clusters of arbitrary shape in noisy spatial data sets.
The steps of the DBSCAN algorithm are as follows: randomly selecting a point p from the point cloud data; if the selected point p is a core point for the parameter neighborhood radius and the number threshold of the points in the neighborhood, finding out all the points with the density reaching from the point p to form a cluster; if the selected point p is an edge point, another point is selected; and repeating the steps until all the points in the point cloud data are processed.
According to the AGV backward pallet and obstacle identification method provided by the embodiment of the invention, the obtained pallet point cloud data can be processed through a normal vector estimation algorithm, a region generation algorithm or a point cloud clustering algorithm, and the position and direction attitude information of the pallet is calculated, so that the AGV can automatically load and unload the pallet.
On the basis of the above embodiment, the AGV backward pallet and obstacle identification method provided in the embodiment of the present invention specifically includes the following steps:
and processing the obstacle point cloud data based on an obstacle perception algorithm or an Euclidean clustering algorithm, and identifying the state information of the obstacle behind the AGV.
Specifically, in the embodiment of the present invention, the obstacle sensing algorithm may be a Normalized Cut (Ncut) algorithm. The basic idea of the Ncut algorithm is based on the assumption of spatial smoothness, i.e. that spatially close points come from the same obstacle. Firstly, the point cloud is preprocessed by using the map information, for example, the point cloud outside the region of interest is removed, and the complexity of the point cloud image is reduced. And then constructing a weighted graph G (V, E, W) according to the preprocessed point clouds, converting the point cloud segmentation into a graph segmentation problem, and solving by utilizing a graph clustering algorithm, wherein each finally solved cluster represents an obstacle.
In the embodiment of the invention, the Euclidean clustering algorithm is a clustering algorithm based on Euclidean distance measurement, and point cloud data smaller than a distance threshold value is taken as a class only according to the distance. The Euclidean clustering algorithm comprises the following steps: finding a certain point p in the point cloud data10Finding the departure point p by using a kd-Tree search algorithm10The nearest n points are judged from the n points to the point p10The distance of (d); the point p with the distance less than the set threshold value12、p13、p14Put in set Q; when the number of elements in Q is not increased any more, clustering is finished, otherwise, p needs to be selected from the set Q10At other points, the above steps are repeated until the elements in the set Q are no longer added. The obstacle can be processed through the Euclidean clustering algorithm, and the state information of the obstacle is identified.
In the embodiment of the present invention, the obstacle state information may include information such as the position, number, and movement of the obstacle.
According to the AGV backward pallet and the obstacle identification method provided by the embodiment of the invention, through the obstacle perception algorithm or the European clustering algorithm, when the pallet point cloud data is used for identifying the AGV backward pallet, the obstacle point cloud data is processed, so that the obstacle state information backward to the AGV can be identified, the AGV can timely avoid obstacles or timely stop when running backward, and unnecessary loss caused by the fact that the AGV touches the obstacles is avoided.
According to the AGV backward pallet and obstacle identification method provided by the embodiment of the invention, the point cloud data acquired by the perception sensor is acquired and processed, the pallet point cloud data and the obstacle point cloud data in the point cloud data are determined, and then the pallet point cloud data and the obstacle point cloud data are respectively identified, so that the position and attitude information of the pallet and the state information of the obstacle are obtained. According to the method, the pallet information and the obstacle information can be acquired simultaneously only by one sensing sensor, and the sensor density of the AGV is reduced.
Fig. 2 is a schematic diagram illustrating a complete flow of an AGV backward pallet and obstacle identification method according to an embodiment of the present invention. As shown in fig. 2, the method includes:
and S21, acquiring point cloud data acquired by the perception sensor.
And S22, processing the point cloud data, and filtering the ground point cloud data. The method for filtering ground point cloud data comprises a random sampling consistency algorithm, and the point cloud data can be preprocessed by filtering, outlier filtering and the like before being filtered out of the ground in order to achieve a better effect.
S23, determining the pallet point cloud data and the obstacle point cloud data, that is, performing the partition of the pallet point cloud and the obstacle point cloud. The method comprises the steps of dividing pallet point cloud data according to at least one of pallet color, target characteristics or geometric characteristics to obtain residual point cloud data, and clustering the residual point cloud data to obtain obstacle point cloud data.
S24, determining whether the pallet point cloud data is segmented, that is, whether the segmentation is successful, and if the segmentation is successful, continuing to execute S25; if the division fails, execution continues with S26.
And S25, respectively identifying the pallet point cloud data and the obstacle point cloud data to obtain the position and attitude information of the pallet and the obstacle state information.
S26, the entire point cloud data is determined as an obstacle, and the obstacle is identified and its state information is calculated.
Fig. 3 is a schematic structural diagram of an AGV backward pallet and obstacle recognition device according to an embodiment of the present invention. As shown in fig. 3, the apparatus includes: an acquisition module 301, a processing module 302 and a recognition module 303. Wherein the content of the first and second substances,
the acquisition module 301 is used for acquiring point cloud data acquired by the perception sensor and backward from the AGV;
the processing module 302 is configured to process the point cloud data, and determine pallet point cloud data and obstacle point cloud data in the point cloud data;
identification module 303 is used for based on the pallet point cloud data discerns the pallet after AGV, and based on obstacle point cloud data discerns the obstacle after AGV.
Specifically, the device for identifying the backward pallet and the obstacle of the AGV in the embodiment of the present invention may be an AGV main controller. The point cloud data collected by the perception sensor can be transmitted to the acquisition module.
In the embodiment of the present invention, the processing of the point cloud data in the processing module may include basic operations such as image enhancement, filtering, key point/edge detection, and the like. After the basic operation is completed, the point cloud data can be segmented, and the pallet point cloud data and the obstacle point cloud data in the point cloud data are determined through segmentation.
In the embodiment of the present invention, the recognition module may recognize the pallet point cloud data obtained based on the segmentation, and the recognition of the pallet may include calculating the direction and position of the pallet, and the like. Meanwhile, based on the obtained obstacle point cloud data, the obstacle can be identified, and the identification of the obstacle can include calculating the position, the number, whether to move and the like of the obstacle.
According to the AGV backward pallet and obstacle recognition device provided by the embodiment of the invention, recognition of the AGV backward pallet and the obstacle can be simultaneously completed only by installing the perception sensor on the AGV, so that the production cost of the AGV is reduced.
On the basis of the above embodiment, the device for identifying a pallet and an obstacle backward for an AGV according to an embodiment of the present invention includes:
the filtering submodule is used for filtering ground point cloud data in the point cloud data to obtain alternative point cloud data;
and the partitioning submodule is used for partitioning the alternative point cloud data to obtain the pallet point cloud data and the obstacle point cloud data.
On the basis of the above embodiment, in the AGV backward pallet and obstacle recognition apparatus provided in the embodiment of the present invention, the filtering submodule is specifically configured to:
performing direct filtering on the point cloud data to determine point cloud data of the region of interest;
and filtering ground point cloud data in the point cloud data of the region of interest to obtain the alternative point cloud data.
On the basis of the above embodiment, in the AGV backward pallet and obstacle recognition apparatus provided in the embodiment of the present invention, the partitioning submodule is specifically configured to:
dividing the pallet point cloud data from the alternative point cloud data based on at least one of pallet color characteristics, pallet target object characteristics and pallet geometric characteristics, and obtaining residual point cloud data in the alternative point cloud data except the pallet point cloud data;
and determining obstacle point cloud data in the residual point cloud data based on a clustering algorithm.
On the basis of the above embodiment, in the device for identifying the backward pallet and the obstacle of the AGV provided in the embodiment of the present invention, the identification module is specifically configured to:
and processing the pallet point cloud data based on a normal vector estimation algorithm, a region generation algorithm or a point cloud clustering algorithm, and identifying the pose information of the AGV to the pallet.
On the basis of the above embodiment, in the AGV backward pallet and obstacle recognition apparatus provided in the embodiment of the present invention, the recognition module is further specifically configured to:
and processing the obstacle point cloud data based on an obstacle perception algorithm or an Euclidean clustering algorithm, and identifying the state information of the obstacle behind the AGV.
Specifically, the functions of the modules in the AGV backward pallet and obstacle recognition device provided in the embodiment of the present invention correspond to the operation flows of the steps in the above method embodiments one to one, and the achieved effects are also consistent, for which, reference is specifically made to the above method embodiments, which are not specifically limited in the embodiment of the present invention.
In a preferred embodiment of the present invention, the AGV backward pallet and obstacle identification device may further include a configuration module, configured to perform parameter configuration on an algorithm in the processing module and the identification module provided in the embodiment of the present invention. The configuration module can set various parameters well first, and the efficiency of the subsequent algorithm is improved.
In a preferred embodiment of the present invention, the device for identifying the pallet and the obstacle after the AGV may further include a User Interface (UI) display module, configured to enable an operator to configure parameters in the configuration module and display a position and an obstacle of the pallet on line. Through the UI display module, the configuration parameters of operators can be conveniently configured, the operators can be enabled to know the pose of the pallet and the state information of the obstacles more clearly, and the AGV can be further conveniently operated.
FIG. 4 is a schematic diagram of an AGV according to an embodiment of the present invention. As shown in FIG. 4, the AGV includes:
perception sensor 401 and as above-mentioned AGV to pallet and obstacle recognition device 402, perception sensor 401 with the AGV is to pallet and obstacle recognition device 402 electricity connection.
Specifically, the AGV provided by the embodiment of the present invention can be applied to the method for identifying the rear pallet and the obstacle of the AGV provided by each embodiment. Therefore, the AGV provided by the embodiment of the present invention has all the advantages of the method for identifying a rear pallet and an obstacle provided by any one of the foregoing embodiments, and details are not repeated herein.
Fig. 5 illustrates a physical structure diagram of an electronic device, which may include, as shown in fig. 5: a processor (processor)510, a communication Interface (Communications Interface)520, a memory (memory)530 and a communication bus 540, wherein the processor 510, the communication Interface 520 and the memory 530 communicate with each other via the communication bus 540. The processor 510 may invoke logic instructions in the memory 530 to perform the AGV backward pallet and obstacle identification method provided in the embodiments described above, the method comprising: acquiring point cloud data behind the AGV, which is acquired by a perception sensor; processing the point cloud data, and determining pallet point cloud data and obstacle point cloud data in the point cloud data; based on the pallet point cloud data is discerned to the pallet after AGV, and based on obstacle point cloud data is discerned to the obstacle after AGV.
Furthermore, the logic instructions in the memory 530 may be implemented in the form of software functional units and stored in a computer readable storage medium when the software functional units are sold or used as independent products. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
In another aspect, the present invention also provides a computer program product comprising a computer program stored on a non-transitory computer readable storage medium, the computer program comprising program instructions which, when executed by a computer, enable the computer to perform the AGV backward pallet and obstacle identification method provided in the above embodiments, the method comprising: acquiring point cloud data behind the AGV, which is acquired by a perception sensor; processing the point cloud data, and determining pallet point cloud data and obstacle point cloud data in the point cloud data; based on the pallet point cloud data is discerned to the pallet after AGV, and based on obstacle point cloud data is discerned to the obstacle after AGV.
In yet another aspect, the present invention further provides a non-transitory computer readable storage medium having a computer program stored thereon, the computer program being implemented by a processor to perform the AGV backward pallet and obstacle identifying method provided in the above embodiments, the method comprising: acquiring point cloud data behind the AGV, which is acquired by a perception sensor; processing the point cloud data, and determining pallet point cloud data and obstacle point cloud data in the point cloud data; based on the pallet point cloud data is discerned to the pallet after AGV, and based on obstacle point cloud data is discerned to the obstacle after AGV.
The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (10)
1. A method for identifying a pallet and a barrier after AGV (automatic guided vehicle) is characterized by comprising the following steps:
acquiring point cloud data behind the AGV, which is acquired by a perception sensor;
processing the point cloud data, and determining pallet point cloud data and obstacle point cloud data in the point cloud data;
based on the pallet point cloud data is discerned to the pallet after AGV, and based on obstacle point cloud data is discerned to the obstacle after AGV.
2. The AGV backward pallet and obstacle identifying method according to claim 1, wherein the point cloud data is processed to determine pallet point cloud data and obstacle point cloud data in the point cloud data, and specifically includes:
filtering ground point cloud data in the point cloud data to obtain alternative point cloud data;
and partitioning the alternative point cloud data to obtain the pallet point cloud data and the obstacle point cloud data.
3. The AGV backward pallet and obstacle identifying method according to claim 2, wherein the ground point cloud data in the point cloud data are filtered out to obtain alternative point cloud data, and the method specifically includes:
performing direct filtering on the point cloud data to determine point cloud data of the region of interest;
and filtering ground point cloud data in the point cloud data of the region of interest to obtain the alternative point cloud data.
4. The AGV backward pallet and obstacle identifying method according to claim 2, wherein the candidate point cloud data is segmented to obtain the pallet point cloud data and the obstacle point cloud data, and the method specifically includes:
dividing the pallet point cloud data from the alternative point cloud data based on at least one of pallet color characteristics, pallet target object characteristics and pallet geometric characteristics, and obtaining residual point cloud data in the alternative point cloud data except the pallet point cloud data;
and determining obstacle point cloud data in the residual point cloud data based on a clustering algorithm.
5. The AGV backward pallet and obstacle identifying method according to any one of claims 1 to 4, wherein the recognition of the AGV backward pallet based on the pallet point cloud data specifically comprises:
and processing the pallet point cloud data based on a normal vector estimation algorithm, a region generation algorithm or a point cloud clustering algorithm, and identifying the pose information of the AGV to the pallet.
6. The AGV backward pallet and obstacle identifying method according to any one of claims 1 to 4, wherein the identifying of the AGV backward obstacle based on the obstacle point cloud data specifically comprises:
and processing the obstacle point cloud data based on an obstacle perception algorithm or an Euclidean clustering algorithm, and identifying the state information of the obstacle behind the AGV.
7. The utility model provides a pallet and barrier recognition device behind AGV which characterized in that includes:
the acquisition module is used for acquiring point cloud data behind the AGV, which is acquired by the perception sensor;
the processing module is used for processing the point cloud data and determining pallet point cloud data and obstacle point cloud data in the point cloud data;
and the recognition module is used for recognizing the pallet after the AGV is processed by the pallet point cloud data and recognizing the obstacle after the AGV is processed by the obstacle point cloud data.
8. An AGV, comprising: the AGV includes a sensing sensor and the device for identifying a pallet and an obstacle according to claim 7, wherein the sensing sensor is electrically connected to the AGV and the pallet and the obstacle identifying device.
9. An electronic device comprising a memory, a processor and a computer program stored on said memory and executable on said processor, wherein said processor when executing said program performs the steps of a method for AGV rear pallet and obstacle identification according to any of claims 1 to 6.
10. A non-transitory computer readable storage medium having stored thereon a computer program, wherein the computer program when executed by a processor implements the steps of an AGV backward pallet and obstacle identification method according to any of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110139701.3A CN112784799A (en) | 2021-02-01 | 2021-02-01 | AGV (automatic guided vehicle) backward pallet and obstacle identification method and device and AGV |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110139701.3A CN112784799A (en) | 2021-02-01 | 2021-02-01 | AGV (automatic guided vehicle) backward pallet and obstacle identification method and device and AGV |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112784799A true CN112784799A (en) | 2021-05-11 |
Family
ID=75760420
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110139701.3A Pending CN112784799A (en) | 2021-02-01 | 2021-02-01 | AGV (automatic guided vehicle) backward pallet and obstacle identification method and device and AGV |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112784799A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113589323A (en) * | 2021-07-07 | 2021-11-02 | 浙江理工大学 | Road obstacle identification method based on height vector field |
| CN115431695A (en) * | 2022-09-09 | 2022-12-06 | 中国第一汽车股份有限公司 | Suspension parameter adjusting method and device, electronic equipment and storage medium |
| CN115641462A (en) * | 2022-12-26 | 2023-01-24 | 电子科技大学 | Radar image target identification method |
| CN116482711A (en) * | 2023-06-21 | 2023-07-25 | 之江实验室 | Local static environment sensing method and device for autonomous selection of landing zone |
| WO2024051054A1 (en) * | 2022-09-08 | 2024-03-14 | 劢微机器人科技(深圳)有限公司 | Transportation obstacle avoidance method, apparatus and device for agv, and storage medium |
Citations (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103955920A (en) * | 2014-04-14 | 2014-07-30 | 桂林电子科技大学 | Binocular vision obstacle detection method based on three-dimensional point cloud segmentation |
| CN107507167A (en) * | 2017-07-25 | 2017-12-22 | 上海交通大学 | A kind of cargo pallet detection method and system matched based on a cloud face profile |
| US20180089517A1 (en) * | 2016-08-10 | 2018-03-29 | Barry D. Douglas | Pallet localization systems and methods |
| CN108931253A (en) * | 2018-07-24 | 2018-12-04 | 福勤智能科技（昆山）有限公司 | Air navigation aid, device, intelligently guiding vehicle and the medium of intelligently guiding vehicle |
| CN109087345A (en) * | 2018-09-06 | 2018-12-25 | 上海仙知机器人科技有限公司 | Pallet recognition methods and automated guided vehicle based on ToF imaging system |
| CN110780276A (en) * | 2019-10-29 | 2020-02-11 | 杭州易博特科技有限公司 | Tray identification method and system based on laser radar and electronic equipment |
| CN110793512A (en) * | 2019-09-11 | 2020-02-14 | 上海宾通智能科技有限公司 | Pose recognition method and device, electronic equipment and storage medium |
| CN110889350A (en) * | 2019-11-18 | 2020-03-17 | 四川西南交大铁路发展股份有限公司 | Line obstacle monitoring and alarming system and method based on three-dimensional imaging |
| CN110889828A (en) * | 2019-11-07 | 2020-03-17 | 浙江大华技术股份有限公司 | Pallet identification method in predetermined scene, terminal device and computer storage medium |
| CN111160302A (en) * | 2019-12-31 | 2020-05-15 | 深圳一清创新科技有限公司 | Obstacle information identification method and device based on automatic driving environment |
| CN111289998A (en) * | 2020-02-05 | 2020-06-16 | 北京汽车集团有限公司 | Obstacle detection method, obstacle detection device, storage medium, and vehicle |
| CN111308500A (en) * | 2020-04-07 | 2020-06-19 | 三一机器人科技有限公司 | Obstacle sensing method and device based on single-line laser radar and computer terminal |
| CN111483949A (en) * | 2020-03-16 | 2020-08-04 | 浙江大华机器人技术有限公司 | Stacking vehicle AGV and height positioning method, device and equipment for pallet fork of stacking vehicle AGV |
| US20200273138A1 (en) * | 2019-02-22 | 2020-08-27 | Dexterity, Inc. | Multicamera image processing |
| CN111681282A (en) * | 2020-06-18 | 2020-09-18 | 浙江大华技术股份有限公司 | Pallet identification processing method and device |
| CN211920797U (en) * | 2019-12-25 | 2020-11-13 | 苏州罗伯特木牛流马物流技术有限公司 | Double-sensor type pallet fork and forklift |
| CN111930125A (en) * | 2020-08-19 | 2020-11-13 | 山东华锐智能技术有限公司 | Low-cost obstacle detection device and method suitable for AGV |
| CN111985322A (en) * | 2020-07-14 | 2020-11-24 | 西安理工大学 | Road environment element sensing method based on laser radar |
-
2021
- 2021-02-01 CN CN202110139701.3A patent/CN112784799A/en active Pending
Patent Citations (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103955920A (en) * | 2014-04-14 | 2014-07-30 | 桂林电子科技大学 | Binocular vision obstacle detection method based on three-dimensional point cloud segmentation |
| US20180089517A1 (en) * | 2016-08-10 | 2018-03-29 | Barry D. Douglas | Pallet localization systems and methods |
| CN107507167A (en) * | 2017-07-25 | 2017-12-22 | 上海交通大学 | A kind of cargo pallet detection method and system matched based on a cloud face profile |
| CN108931253A (en) * | 2018-07-24 | 2018-12-04 | 福勤智能科技（昆山）有限公司 | Air navigation aid, device, intelligently guiding vehicle and the medium of intelligently guiding vehicle |
| CN109087345A (en) * | 2018-09-06 | 2018-12-25 | 上海仙知机器人科技有限公司 | Pallet recognition methods and automated guided vehicle based on ToF imaging system |
| US20200273138A1 (en) * | 2019-02-22 | 2020-08-27 | Dexterity, Inc. | Multicamera image processing |
| CN110793512A (en) * | 2019-09-11 | 2020-02-14 | 上海宾通智能科技有限公司 | Pose recognition method and device, electronic equipment and storage medium |
| CN110780276A (en) * | 2019-10-29 | 2020-02-11 | 杭州易博特科技有限公司 | Tray identification method and system based on laser radar and electronic equipment |
| CN110889828A (en) * | 2019-11-07 | 2020-03-17 | 浙江大华技术股份有限公司 | Pallet identification method in predetermined scene, terminal device and computer storage medium |
| CN110889350A (en) * | 2019-11-18 | 2020-03-17 | 四川西南交大铁路发展股份有限公司 | Line obstacle monitoring and alarming system and method based on three-dimensional imaging |
| CN211920797U (en) * | 2019-12-25 | 2020-11-13 | 苏州罗伯特木牛流马物流技术有限公司 | Double-sensor type pallet fork and forklift |
| CN111160302A (en) * | 2019-12-31 | 2020-05-15 | 深圳一清创新科技有限公司 | Obstacle information identification method and device based on automatic driving environment |
| CN111289998A (en) * | 2020-02-05 | 2020-06-16 | 北京汽车集团有限公司 | Obstacle detection method, obstacle detection device, storage medium, and vehicle |
| CN111483949A (en) * | 2020-03-16 | 2020-08-04 | 浙江大华机器人技术有限公司 | Stacking vehicle AGV and height positioning method, device and equipment for pallet fork of stacking vehicle AGV |
| CN111308500A (en) * | 2020-04-07 | 2020-06-19 | 三一机器人科技有限公司 | Obstacle sensing method and device based on single-line laser radar and computer terminal |
| CN111681282A (en) * | 2020-06-18 | 2020-09-18 | 浙江大华技术股份有限公司 | Pallet identification processing method and device |
| CN111985322A (en) * | 2020-07-14 | 2020-11-24 | 西安理工大学 | Road environment element sensing method based on laser radar |
| CN111930125A (en) * | 2020-08-19 | 2020-11-13 | 山东华锐智能技术有限公司 | Low-cost obstacle detection device and method suitable for AGV |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113589323A (en) * | 2021-07-07 | 2021-11-02 | 浙江理工大学 | Road obstacle identification method based on height vector field |
| WO2024051054A1 (en) * | 2022-09-08 | 2024-03-14 | 劢微机器人科技(深圳)有限公司 | Transportation obstacle avoidance method, apparatus and device for agv, and storage medium |
| CN115431695A (en) * | 2022-09-09 | 2022-12-06 | 中国第一汽车股份有限公司 | Suspension parameter adjusting method and device, electronic equipment and storage medium |
| CN115641462A (en) * | 2022-12-26 | 2023-01-24 | 电子科技大学 | Radar image target identification method |
| CN116482711A (en) * | 2023-06-21 | 2023-07-25 | 之江实验室 | Local static environment sensing method and device for autonomous selection of landing zone |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zermas et al. | Fast segmentation of 3d point clouds: A paradigm on lidar data for autonomous vehicle applications | |
| CN112784799A (en) | AGV (automatic guided vehicle) backward pallet and obstacle identification method and device and AGV | |
| US10031231B2 (en) | Lidar object detection system for automated vehicles | |
| Nguyen et al. | 3D point cloud segmentation: A survey | |
| Pătrăucean et al. | State of research in automatic as-built modelling | |
| Douillard et al. | On the segmentation of 3D LIDAR point clouds | |
| Hackel et al. | Joint classification and contour extraction of large 3D point clouds | |
| WO2022188663A1 (en) | Target detection method and apparatus | |
| US20220222824A1 (en) | Fully automated multimodal system architecture for semantic segmentation of large-scale 3d outdoor point cloud data | |
| CN112529015A (en) | Three-dimensional point cloud processing method, device and equipment based on geometric unwrapping | |
| Drews et al. | Novelty detection and 3d shape retrieval using superquadrics and multi-scale sampling for autonomous mobile robots | |
| Son et al. | Rapid 3D object detection and modeling using range data from 3D range imaging camera for heavy equipment operation | |
| Choe et al. | Fast point cloud segmentation for an intelligent vehicle using sweeping 2D laser scanners | |
| Yogeswaran et al. | 3d surface analysis for automated detection of deformations on automotive body panels | |
| CN113420637A (en) | Laser radar detection method under multi-scale aerial view angle in automatic driving | |
| Luperto et al. | Extracting structure of buildings using layout reconstruction | |
| Kaushik et al. | Accelerated patch-based planar clustering of noisy range images in indoor environments for robot mapping | |
| CN114519681A (en) | Automatic calibration method and device, computer readable storage medium and terminal | |
| Hesami et al. | Range segmentation of large building exteriors: A hierarchical robust approach | |
| Keller et al. | Extracting and visualizing structural features in environmental point cloud LiDaR data sets | |
| Kovacs et al. | Edge detection in discretized range images | |
| Kaleci et al. | Plane segmentation of point cloud data using split and merge based method | |
| CN115147564A (en) | Three-dimensional model construction method, neural network training method and device | |
| Truax et al. | Using prioritized relaxations to locate objects in points clouds for manipulation | |
| Specht et al. | Edge registration versus triangular mesh registration, a comparative study |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |