CN114200946B - AGV trolley control method for intelligent manufacturing machining production line - Google Patents
AGV trolley control method for intelligent manufacturing machining production line Download PDFInfo
- Publication number
- CN114200946B CN114200946B CN202111524868.8A CN202111524868A CN114200946B CN 114200946 B CN114200946 B CN 114200946B CN 202111524868 A CN202111524868 A CN 202111524868A CN 114200946 B CN114200946 B CN 114200946B
- Authority
- CN
- China
- Prior art keywords
- detection
- shadow
- points
- judging
- image
- 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.)
- Active
Links
- 238000004519 manufacturing process Methods 0.000 title claims abstract description 21
- 238000000034 method Methods 0.000 title claims abstract description 19
- 238000003754 machining Methods 0.000 title claims abstract description 10
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 10
- 238000000605 extraction Methods 0.000 claims abstract description 8
- 238000001514 detection method Methods 0.000 claims description 60
- 238000010586 diagram Methods 0.000 claims description 13
- 238000005259 measurement Methods 0.000 claims description 2
- 238000004364 calculation method Methods 0.000 abstract description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0231—Control of position or course in two dimensions specially adapted to land vehicles using optical position detecting means
- G05D1/0246—Control of position or course in two dimensions specially adapted to land vehicles using optical position detecting means using a video camera in combination with image processing means
- G05D1/0253—Control of position or course in two dimensions specially adapted to land vehicles using optical position detecting means using a video camera in combination with image processing means extracting relative motion information from a plurality of images taken successively, e.g. visual odometry, optical flow
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Electromagnetism (AREA)
- Aviation & Aerospace Engineering (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses an AGV trolley control method for an intelligent manufacturing and machining production line, which comprises the following steps: acquiring an original image output by a camera module; pre-judging an original image by using gray processing; the pre-judging result is sent to the control module to correspondingly control the driving module; re-judging the image by using a convolutional neural network; and sending the re-judging result to the control module to correspondingly control the driving module. The original image is processed and judged sequentially by utilizing gray processing and convolutional neural network processing, and the result is sent to the control module to control the driving module so as to achieve the purpose of accurately controlling the AGV trolley, thereby solving the problem that the AGV trolley cannot be effectively identified when sundry obstacles such as screw wrenches falling near a processing production line appear on a planned track. Meanwhile, the convolutional neural network processing directly acquires the characteristic image of gray level processing as the characteristic extraction image to process, so that the time of calculation processing is greatly shortened, and the timeliness of the scheme is improved.
Description
Technical Field
The invention is applied to the field of AGV trolley control, and particularly relates to an AGV trolley control method for an intelligent manufacturing and machining production line.
Background
The AGV car is a transport vehicle equipped with an automatic navigation device such as electromagnetic or optical, capable of traveling along a predetermined navigation path, and having safety protection and various transfer functions. In industrial application, a carrier for a driver is not needed, and a rechargeable storage battery is used as a power source. Generally, the traveling path and behavior of the vehicle can be controlled by a computer, or the traveling path can be established by using an electromagnetic track (electromagnetic path-following system), the electromagnetic track is stuck on the floor, and the vehicle is moved and operated by the information brought by the electromagnetic track.
Most of the existing AGV trolleys adopt electromagnetic tracks or are preset to be changed into tracks for navigation, but the corresponding obstacle avoidance capability is lacking, when sundry obstacles such as screw wrenches falling near a processing production line appear on a planned track, the AGV trolleys cannot be effectively identified, and therefore how to enable the AGV trolleys to effectively identify the obstacles on a forward path is a key technical problem to be solved in the field.
Disclosure of Invention
The invention aims to solve the technical problem of providing an AGV trolley control method for an intelligent manufacturing and machining production line, aiming at the defects of the prior art.
In order to solve the technical problems, the AGV trolley control method for the intelligent manufacturing and machining production line comprises a camera module for shooting the forward road condition of the AGV trolley, a driving module for driving the AGV trolley to move and a control module for controlling the driving module to work; the AGV trolley control method comprises the following steps:
Acquiring an original image output by a camera module;
Pre-judging an original image by using gray processing;
the pre-judging result is sent to the control module to correspondingly control the driving module;
Re-judging the image by using a convolutional neural network;
and sending the re-judging result to the control module to correspondingly control the driving module.
As a possible implementation manner, the pre-judging step of the original image by using gray processing further includes:
acquiring a characteristic region diagram;
Obtaining a shadow pattern;
Obtaining an object graph;
and comparing and judging the shadow pattern with the object pattern.
As a possible implementation manner, the step of acquiring the feature area map specifically includes:
performing traversal detection on the gray value of the acquired original image according to preset detection points to obtain the gray value of each detection point;
marking the adjacent detection points with the gray value difference reaching the preset difference, and performing rectangular frame selection on the marked detection points and intercepting the marked detection points as a characteristic area diagram.
As a possible implementation manner, further, the step of obtaining the shadow pattern specifically includes:
performing traversal detection on gray values of the characteristic area graph according to preset detection points to obtain the gray values of the detection points;
Marking a detection point with a gray value within a preset range as a shadow point, and connecting the marked shadow point as a closed graph;
And measuring the length and width of the two ends of the closed graph connected by the shadow points.
As a possible implementation manner, the step of obtaining the object graph specifically includes:
performing traversal detection on gray values of the characteristic area graph according to preset detection points to obtain the gray values of the detection points;
Marking detection points with gray values within a preset range as object points, and connecting the marked object points into a closed graph;
And measuring the length and width of the two ends of the closed graph connected with the object points.
As a possible implementation manner, the step of comparing and judging the shadow pattern with the object pattern specifically includes: and carrying out rationality comparison on the closed graph with the shadow points and the object points connected, and outputting a result of comparison judgment on rationality or not.
As a possible implementation manner, the sending the pre-judging result to the control module to perform the corresponding control step on the driving module specifically includes:
The control module receives a pre-judging result of whether the shadow pattern is reasonably matched with the object pattern;
If the pre-judging result is reasonable match, judging that the object in the object graph is a real object, and sending a creep or avoiding instruction to a driving module by the control module;
If the pre-judging result is unreasonable matching, judging that the object in the object graph is non-physical, and sending a normal running instruction to the driving module by the control module.
As a possible implementation manner, the step of re-judging the image by using the convolutional neural network specifically includes:
taking the characteristic region diagram intercepted in the step of acquiring the characteristic region diagram as a characteristic extraction image;
and extracting the object and shadow in the image by utilizing the RPN network discrimination characteristics, and carrying out regression processing on the position of the object.
As a possible implementation manner, the step of extracting the object and the shadow in the image by using the RPN network discrimination feature and performing regression processing on the position of the object specifically includes:
Performing convolution operation on the feature extraction image, forming a plurality of anchor points by taking pixel points in the feature image as centers, and performing candidate frame selection operation by using a preset rectangular frame;
the classification layer in the RPN judges the candidate frames and obtains the scores of whether the objects or shadows in the candidate frames;
And the regression layer in the RPN network layer carries out regression operation on the total number of the selected candidate frames to extract the position information of the object and the shadow in the image.
As one possible implementation, the device may, further,
Marking the detection point with the gray value within the preset range as a shadow point, wherein the preset range in the step of marking the detection point with the gray value within the preset range is that the gray value is between 0 and 110;
And marking the detection point with the gray value within the preset range as an object point, wherein the preset range in the step is that the gray value is between 110 and 255.
The invention adopts the technical scheme and has the following beneficial effects:
According to the invention, the original image is processed and judged sequentially by utilizing gray processing and convolutional neural network processing, and the result is sent to the control module to control the driving module so as to achieve the purpose of accurately controlling the AGV, thereby solving the problem that the AGV cannot be effectively identified when the planning track is provided with sundry obstacles such as screw wrenches falling near a processing production line. Meanwhile, the convolutional neural network processing directly acquires the characteristic image of gray level processing as the characteristic extraction image to process, so that the time of calculation processing is greatly shortened, and the timeliness of the scheme is improved.
Drawings
The invention is described in further detail below with reference to the attached drawings and detailed description:
FIG. 1 is a schematic flow diagram of the principles of the present invention;
FIG. 2 is a schematic view of a characteristic region acquisition interface of the present invention;
FIG. 3 is a schematic diagram of a shadow graphic and object graphic judgment interface according to the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.
As shown in fig. 1, the invention provides an intelligent manufacturing and machining production line AGV trolley control method, wherein the AGV trolley comprises a camera module for shooting the forward road condition of the AGV trolley, a driving module for driving the AGV trolley to move and a control module for controlling the driving module to work; the AGV trolley control method comprises the following steps:
Acquiring an original image output by a camera module;
Pre-judging an original image by using gray processing; comprising the following steps: acquiring a characteristic region map as shown in fig. 2; further, the step of obtaining the feature area map specifically includes: performing traversal detection on the gray value of the acquired original image according to preset detection points to obtain the gray value of each detection point; marking the adjacent detection points with the gray value difference reaching the preset difference, and performing quadrilateral frame selection on the marked detection points and intercepting the marked detection points as a characteristic area diagram. Obtaining a shadow pattern; further, the step of obtaining the shadow pattern specifically includes: performing traversal detection on gray values of the characteristic area graph according to preset detection points to obtain the gray values of the detection points; marking a detection point with a gray value within a preset range as a shadow point, and connecting the marked shadow point as a closed graph; and measuring the length and width of the two ends of the closed graph connected by the shadow points. Obtaining an object graph; the step of obtaining the object graph comprises the following steps: performing traversal detection on gray values of the characteristic area graph according to preset detection points to obtain the gray values of the detection points; marking detection points with gray values within a preset range as object points, and connecting the marked object points into a closed graph; and measuring the length and width of the two ends of the closed graph connected with the object points. And comparing and judging the shadow pattern with the object pattern. The method comprises the following steps: and carrying out rationality comparison on the closed graph with the shadow points and the object points connected, and outputting a result of comparison judgment on rationality or not.
The light source direction combination judgment scheme comprises the following steps:
(1) Circumferentially surrounding the outer wall of the AGV trolley body and arranging light source detection equipment so as to detect the real-time light source direction of the position of the trolley and upload detection data in real time;
(2) Acquiring real-time light source direction detection data, and combining the real-time light source direction detection data with relative position data and dimension measurement data of a shadow pattern and an object pattern to compare the position dimension rationality of the shadow pattern and the object pattern;
(3) And sending the rationality comparison result to the control module to correspondingly control the driving module.
As one possible implementation, the device may, further,
Marking the detection point with the gray value within the preset range as a shadow point, wherein the preset range in the step of marking the detection point with the gray value within the preset range is that the gray value is between 0 and 110;
And marking the detection point with the gray value within the preset range as an object point, wherein the preset range in the step is that the gray value is between 110 and 255.
The pre-judging result is sent to the control module to correspondingly control the driving module; the method specifically comprises the following steps:
The control module receives a pre-judging result of whether the shadow pattern is reasonably matched with the object pattern;
If the pre-judging result is reasonable match, judging that the object in the object graph is a real object, and sending a creep or avoiding instruction to a driving module by the control module;
If the pre-judging result is unreasonable matching, judging that the object in the object graph is non-physical, and sending a normal running instruction to the driving module by the control module.
Re-judging the image by using a convolutional neural network; the method specifically comprises the following steps:
taking the characteristic region diagram intercepted in the step of acquiring the characteristic region diagram as a characteristic extraction image;
And extracting the object and shadow in the image by utilizing the RPN network discrimination characteristics as shown in fig. 3, and carrying out regression processing on the position of the object.
As a possible implementation manner, the step of extracting the object and the shadow in the image by using the RPN network discrimination feature and performing regression processing on the position of the object specifically includes:
Performing convolution operation on the feature extraction image, forming a plurality of anchor points by taking pixel points in the feature image as centers, and performing candidate frame selection operation by using a preset rectangular frame;
the classification layer in the RPN judges the candidate frames and obtains the scores of whether the objects or shadows in the candidate frames;
And the regression layer in the RPN network layer carries out regression operation on the total number of the selected candidate frames to extract the position information of the object and the shadow in the image.
And sending the re-judging result to the control module to correspondingly control the driving module.
While the foregoing is directed to embodiments of the present invention, other and further embodiments of the invention may be devised without departing from the basic scope thereof, and the scope thereof is determined by the claims that follow.
Claims (3)
1. The AGV trolley control method for the intelligent manufacturing machining production line is characterized by comprising a camera module for shooting the forward road condition of the AGV trolley, a driving module for driving the AGV trolley to move and a control module for controlling the driving module to work; the AGV trolley control method comprises the following steps:
Acquiring an original image output by a camera module;
pre-judging an original image by using gray processing; the method comprises the following steps: acquiring a characteristic region diagram; obtaining a shadow pattern; obtaining an object graph; comparing and judging the shadow pattern and the object pattern;
the step of acquiring the characteristic region map specifically comprises the following steps:
performing traversal detection on the gray value of the acquired original image according to preset detection points to obtain the gray value of each detection point;
marking adjacent detection points with gray value difference reaching a preset difference, and carrying out rectangular frame selection on the marked detection points and intercepting the marked detection points as a characteristic area diagram;
The shadow pattern obtaining step specifically comprises the following steps:
performing traversal detection on gray values of the characteristic area graph according to preset detection points to obtain the gray values of the detection points;
Marking a detection point with a gray value within a preset range as a shadow point, and connecting the marked shadow point as a closed graph;
measuring the length and width of the two ends of the closed graph connected with the shadow points;
The step of obtaining the object graph comprises the following steps:
performing traversal detection on gray values of the characteristic area graph according to preset detection points to obtain the gray values of the detection points;
Marking detection points with gray values within a preset range as object points, and connecting the marked object points into a closed graph;
measuring the length and width of the two ends of the closed graph connected with the object points;
The step of comparing and judging the shadow pattern and the object pattern specifically comprises the following steps: carrying out rationality comparison on the closed graph connected by the shadow points and the object points, and outputting a result of comparison judgment on rationality or not;
the light source direction combination judgment scheme comprises the following steps: (1) Circumferentially surrounding the outer wall of the AGV trolley body and arranging light source detection equipment so as to detect the real-time light source direction of the position of the trolley and upload detection data in real time; (2) Acquiring real-time light source direction detection data, and combining the real-time light source direction detection data with relative position data and dimension measurement data of a shadow pattern and an object pattern to compare the position dimension rationality of the shadow pattern and the object pattern; (3) Sending the rationality comparison result to a control module to correspondingly control the driving module;
the pre-judging result is sent to the control module to correspondingly control the driving module;
Re-judging the image by using a convolutional neural network;
The re-judging result is sent to the control module to correspondingly control the driving module;
the step of re-judging the image by using the convolutional neural network specifically comprises the following steps:
taking the characteristic region diagram intercepted in the step of acquiring the characteristic region diagram as a characteristic extraction image;
extracting an object and shadow in the image by utilizing the RPN network discrimination characteristics, and carrying out regression processing on the position of the object;
the step of extracting the object and shadow in the image by utilizing the RPN network discrimination characteristics and carrying out regression processing on the position of the object specifically comprises the following steps:
Performing convolution operation on the feature extraction image, forming a plurality of anchor points by taking pixel points in the feature image as centers, and performing candidate frame selection operation by using a preset rectangular frame;
the classification layer in the RPN judges the candidate frames and obtains the scores of whether the objects or shadows in the candidate frames;
And the regression layer in the RPN network layer carries out regression operation on the total number of the selected candidate frames to extract the position information of the object and the shadow in the image.
2. The intelligent manufacturing and machining line AGV trolley control method according to claim 1 is characterized in that:
The step of sending the pre-judging result to the control module to correspondingly control the driving module specifically comprises the following steps:
The control module receives a pre-judging result of whether the shadow pattern is reasonably matched with the object pattern;
If the pre-judging result is reasonable match, judging that the object in the object graph is a real object, and sending a creep or avoiding instruction to a driving module by the control module;
If the pre-judging result is unreasonable matching, judging that the object in the object graph is non-physical, and sending a normal running instruction to the driving module by the control module.
3. The intelligent manufacturing and machining line AGV trolley control method according to claim 1 is characterized in that:
marking the detection point with the gray value within the preset range as a shadow point, wherein the preset range in the step of marking the detection point with the gray value within the preset range is that the gray value is between 0 and 110;
And marking the detection point with the gray value within the preset range as the object point, wherein the preset range in the step of marking the detection point with the gray value within the preset range is that the gray value is between 110 and 255.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111524868.8A CN114200946B (en) | 2021-12-14 | 2021-12-14 | AGV trolley control method for intelligent manufacturing machining production line |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111524868.8A CN114200946B (en) | 2021-12-14 | 2021-12-14 | AGV trolley control method for intelligent manufacturing machining production line |
Publications (2)
Publication Number | Publication Date |
---|---|
CN114200946A CN114200946A (en) | 2022-03-18 |
CN114200946B true CN114200946B (en) | 2024-05-28 |
Family
ID=80653368
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202111524868.8A Active CN114200946B (en) | 2021-12-14 | 2021-12-14 | AGV trolley control method for intelligent manufacturing machining production line |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114200946B (en) |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112036210A (en) * | 2019-06-03 | 2020-12-04 | 杭州海康机器人技术有限公司 | Method and device for detecting obstacle, storage medium and mobile robot |
CN112163667A (en) * | 2020-09-16 | 2021-01-01 | 闽江学院 | Novel Faster R-CNN network model and training method thereof |
CN112498339A (en) * | 2019-09-13 | 2021-03-16 | 朱宏 | Automatic driving system of motor vehicle |
CN113592911A (en) * | 2021-07-31 | 2021-11-02 | 西南电子技术研究所(中国电子科技集团公司第十研究所) | Apparent enhanced depth target tracking method |
-
2021
- 2021-12-14 CN CN202111524868.8A patent/CN114200946B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112036210A (en) * | 2019-06-03 | 2020-12-04 | 杭州海康机器人技术有限公司 | Method and device for detecting obstacle, storage medium and mobile robot |
CN112498339A (en) * | 2019-09-13 | 2021-03-16 | 朱宏 | Automatic driving system of motor vehicle |
CN112163667A (en) * | 2020-09-16 | 2021-01-01 | 闽江学院 | Novel Faster R-CNN network model and training method thereof |
CN113592911A (en) * | 2021-07-31 | 2021-11-02 | 西南电子技术研究所(中国电子科技集团公司第十研究所) | Apparent enhanced depth target tracking method |
Also Published As
Publication number | Publication date |
---|---|
CN114200946A (en) | 2022-03-18 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112036210B (en) | Method and device for detecting obstacle, storage medium and mobile robot | |
CN109159113B (en) | Robot operation method based on visual reasoning | |
CN109797691A (en) | Unmanned sweeper and its travelling-crane method | |
CN108416320A (en) | Inspection device, the control method of inspection device and control device | |
CN112528979B (en) | Transformer substation inspection robot obstacle distinguishing method and system | |
CN108919810A (en) | The localization for Mobile Robot and navigation system of view-based access control model teaching | |
CN108106617A (en) | A kind of unmanned plane automatic obstacle-avoiding method | |
CN115937213A (en) | Visual defect identification system of automatic inspection robot for mining monorail crane track | |
CN111813114A (en) | Intelligent car visual navigation method | |
CN109343517A (en) | Crusing robot pinpoints visiting method | |
CN113674355A (en) | Target identification and positioning method based on camera and laser radar | |
CN106272434A (en) | One key of a kind of crusing robot makes a return voyage control method and system | |
CN115223039A (en) | Robot semi-autonomous control method and system for complex environment | |
CN107506753A (en) | A kind of more wireless vehicle trackings towards dynamic video monitoring | |
CN114200946B (en) | AGV trolley control method for intelligent manufacturing machining production line | |
CN207516258U (en) | The multifunctional examining measuring car of narrow space | |
CN116922448B (en) | Environment sensing method, device and system for high-speed railway body-in-white transfer robot | |
CN112947461A (en) | VSLAM algorithm-based blast furnace tuyere platform inspection robot control method | |
CN117369460A (en) | Intelligent inspection method and system for loosening faults of vehicle bolts | |
CN109032125B (en) | Navigation method of visual AGV | |
CN117115249A (en) | Container lock hole automatic identification and positioning system and method | |
CN115755888A (en) | AGV obstacle detection system with multi-sensor data fusion and obstacle avoidance method | |
CN116022657A (en) | Path planning method and device and crane | |
CN115661966A (en) | Inspection system and method based on augmented reality | |
CN112015180B (en) | Intelligent experiment trolley and control system thereof |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |