CN111731324A - Control method and system for guiding AGV intelligent vehicle based on vision - Google Patents
Control method and system for guiding AGV intelligent vehicle based on vision Download PDFInfo
- Publication number
- CN111731324A CN111731324A CN202010475734.0A CN202010475734A CN111731324A CN 111731324 A CN111731324 A CN 111731324A CN 202010475734 A CN202010475734 A CN 202010475734A CN 111731324 A CN111731324 A CN 111731324A
- Authority
- CN
- China
- Prior art keywords
- road
- intelligent vehicle
- boundary
- vehicle
- 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.)
- Pending
Links
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W60/00—Drive control systems specially adapted for autonomous road vehicles
- B60W60/001—Planning or execution of driving tasks
- B60W60/0011—Planning or execution of driving tasks involving control alternatives for a single driving scenario, e.g. planning several paths to avoid obstacles
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W30/00—Purposes of road vehicle drive control systems not related to the control of a particular sub-unit, e.g. of systems using conjoint control of vehicle sub-units, or advanced driver assistance systems for ensuring comfort, stability and safety or drive control systems for propelling or retarding the vehicle
- B60W30/08—Active safety systems predicting or avoiding probable or impending collision or attempting to minimise its consequences
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W2552/00—Input parameters relating to infrastructure
- B60W2552/50—Barriers
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W2552/00—Input parameters relating to infrastructure
- B60W2552/53—Road markings, e.g. lane marker or crosswalk
Abstract
The invention discloses a control method and a control system for guiding an AGV (automatic guided vehicle) based on vision, belongs to the technical field of AGV intelligent vehicles, and solves the problem that the tracking precision of the AGV intelligent vehicle in the prior art is low. A control method for guiding an AGV intelligent vehicle based on vision comprises the following steps: acquiring a road image of a road where an intelligent vehicle is located, preprocessing the road image to obtain a preprocessed image, scanning and comparing the preprocessed image to acquire a boundary and a road condition of the road where the intelligent vehicle is located; acquiring road center line information according to the boundary of the road where the intelligent vehicle is located and the road condition, and planning a driving path of the intelligent vehicle according to the center line information; and carrying out weighted summation processing according to the running path to obtain a deviation value of the actual vehicle body and the planned path, and adjusting the running track of the intelligent vehicle according to the running path and the deviation value. The method improves the tracking precision of the AGV intelligent vehicle.
Description
Technical Field
The invention relates to the technical field of AGV intelligent vehicles, in particular to a control method and a system for guiding an AGV intelligent vehicle based on vision.
Background
The AGV includes a transport vehicle equipped with an electromagnetic or optical automatic guide device, capable of traveling along a predetermined guide path, and having safety protection and various transfer functions; the AGV intelligent vehicle is widely applied to the warehousing industry, the manufacturing industry, the post office, the library and the competition field, the tracking program of the existing AGV intelligent vehicle is complex, and the tracking precision and the anti-interference capability of the AGV intelligent vehicle are seriously influenced.
Disclosure of Invention
In view of the above, the invention provides a control method and system for guiding an AGV intelligent vehicle based on vision, which solve the technical problem that the tracking precision of the AGV intelligent vehicle in the prior art is low.
In one aspect, the invention provides a control method for guiding an AGV based on vision, which comprises the following steps:
acquiring a road image of a road where an intelligent vehicle is located, preprocessing the road image to obtain a preprocessed image, scanning and comparing the preprocessed image to acquire a boundary and a road condition of the road where the intelligent vehicle is located;
acquiring road center line information according to the boundary of the road where the intelligent vehicle is located and the road condition, and planning a driving path of the intelligent vehicle according to the center line information;
and carrying out weighted summation processing according to the running path to obtain a deviation value of the actual vehicle body and the planned path, and adjusting the running track of the intelligent vehicle according to the running path and the deviation value.
Further, the road image is preprocessed to obtain a preprocessed image, the preprocessed image is scanned and compared to obtain the boundary and the road condition of the road where the intelligent vehicle is located, and the method specifically comprises the following steps of,
carrying out graying, binarization and compression processing on the road image, separating the outlines of a road and a non-road to obtain a preprocessed image, scanning the middle of the road outline in the preprocessed image to two sides to obtain pixel jumping points to obtain the boundary of the road where the intelligent vehicle is located, and obtaining the road condition of the intelligent vehicle according to the boundary of the road where the intelligent vehicle is located, wherein the road condition is any one of a straight road, a cross road, a roundabout road and a curve road.
Furthermore, road center line information is obtained according to the boundary of the road where the intelligent vehicle is located and the road condition, and the driving path of the intelligent vehicle is planned according to the center line information,
dividing the preset road into n rows and m columns to form an n x m array of the preset road, and pre-storing the width of each row of the preset road in an n x m size graph;
if the road condition of the intelligent vehicle is a straight road, a cross road or a curve road, then
For the boundary of the road, if the left boundary and the right boundary exist, the midpoint of the corresponding line in the road is the average value of the corresponding numerical values of the two boundaries, if the left boundary does not have the right boundary, the midpoint of the corresponding line is obtained by adding half of the width of the road of the corresponding line to the left boundary, if the right boundary does not have the left boundary, the midpoint of the corresponding line is obtained by subtracting half of the width of the road of the corresponding line from the right boundary, if the left boundary and the right boundary are both boundaries, the midpoint of the previous line is taken as the midpoint of the current line, the corresponding midpoints of all the lines are connected to obtain centerline information, the centerline information is the planned driving path of the intelligent vehicle,
if the road condition of the intelligent vehicle is the roundabout, then
And connecting the pixel jumping points with the left boundary point at the most proximal end of the image, taking the connecting line as a left boundary, and then planning the driving path of the intelligent vehicle when the road condition of the intelligent vehicle is a straight road, a cross road or a curve.
And further, carrying out weighted summation processing according to the driving path to obtain a deviation value of the actual vehicle body and the planned path, specifically comprising carrying out amplitude value processing and weighted summation on the center of each row of the preset road, and carrying out difference between the summation result and the road position center line to obtain the deviation value of the actual vehicle body and the planned path.
The method for controlling the AGV intelligent vehicle based on the visual guidance further comprises the steps of detecting whether an obstacle exists, if so, obtaining the index position of the obstacle in an n x m array, determining whether the obstacle is on the left side or the right side of the intelligent vehicle, taking the advancing direction of the intelligent vehicle as the positive direction of an x axis, obtaining the vertical distance from the obstacle to the x axis, and if the ratio of the vertical distance to the half vehicle width of the intelligent vehicle is smaller than a preset value, enabling the intelligent vehicle to avoid the obstacle to the right side or the left side according to whether the obstacle is on the left side or the right side of the intelligent vehicle.
On the other hand, the invention provides a control system for guiding an AGV (automatic guided vehicle) based on vision, which comprises a road state acquisition module, a driving path planning module and a driving track adjusting module;
the road state acquisition module is used for acquiring a road image of a road where the intelligent vehicle is located, preprocessing the road image to obtain a preprocessed image, scanning and comparing the preprocessed image to acquire the boundary and the road condition of the road where the intelligent vehicle is located;
the driving path planning module is used for acquiring road center line information according to the boundary of the road where the intelligent vehicle is located and the road condition, and planning the driving path of the intelligent vehicle according to the center line information;
and the running track adjusting module is used for carrying out weighting and summing processing according to the running path, acquiring a deviation value of an actual vehicle body and the planned path, and adjusting the running track of the intelligent vehicle according to the running path and the deviation value.
Further, the road state obtaining module is used for preprocessing the road image to obtain a preprocessed image, scanning and comparing the preprocessed image to obtain the boundary and the road condition of the road where the intelligent vehicle is located, and specifically comprises,
carrying out graying, binarization and compression processing on the road image, separating the outlines of a road and a non-road to obtain a preprocessed image, scanning the middle of the road outline in the preprocessed image to two sides to obtain pixel jumping points to obtain the boundary of the road where the intelligent vehicle is located, and obtaining the road condition of the intelligent vehicle according to the boundary of the road where the intelligent vehicle is located, wherein the road condition is any one of a straight road, a cross road, a roundabout road and a curve road.
Furthermore, the driving path planning module acquires road center line information according to the boundary of the road where the intelligent vehicle is located and the road condition, and plans the driving path of the intelligent vehicle according to the center line information, and specifically comprises,
dividing the preset road into n rows and m columns to form an n x m array of the preset road, and pre-storing the width of each row of the preset road in an n x m size graph;
if the road condition of the intelligent vehicle is a straight road, a cross road or a curve road, then
For the boundary of the road, if the left boundary and the right boundary exist, the midpoint of the corresponding line in the road is the average value of the corresponding numerical values of the two boundaries, if the left boundary does not have the right boundary, the midpoint of the corresponding line is obtained by adding half of the width of the road of the corresponding line to the left boundary, if the right boundary does not have the left boundary, the midpoint of the corresponding line is obtained by subtracting half of the width of the road of the corresponding line from the right boundary, if the left boundary and the right boundary are both boundaries, the midpoint of the previous line is taken as the midpoint of the current line, the corresponding midpoints of all the lines are connected, and the midline information is the planned driving path of the intelligent vehicle.
If the road condition of the intelligent vehicle is the roundabout, then
And connecting the pixel jumping points with the left boundary point at the most proximal end of the image, taking the connecting line as a left boundary, and then planning the driving path of the intelligent vehicle when the road condition of the intelligent vehicle is a straight road, a cross road or a curve.
The driving track adjusting module is used for carrying out weighted summation processing according to the driving path to obtain a deviation value of the actual vehicle body and the planned path.
Further, control system based on vision guide AGV intelligence car still includes keeps away the barrier module, keep away the barrier module for whether detect and have the barrier, if there is the barrier, then acquire the index position that the barrier is located in n x m's array, confirm the barrier is on the left side or the right side of intelligence car to the direction of advancing is the x axle positive direction before the intelligence car, acquires the barrier to the vertical distance of x axle, if the ratio of vertical distance and the half car width of intelligence car is less than the default, makes the intelligence car according to the barrier is on the left side or the right side of intelligence car, keeps away the barrier to right side or left side.
Compared with the prior art, the invention has the beneficial effects that: the method comprises the steps of preprocessing a road image by acquiring the road image of a road where an intelligent vehicle is located to obtain a preprocessed image, and scanning and comparing the preprocessed image to acquire the boundary and road conditions of the road where the intelligent vehicle is located; acquiring road center line information according to the boundary of the road where the intelligent vehicle is located and the road condition, and planning a driving path of the intelligent vehicle according to the center line information; carrying out weighted summation processing according to the running path to obtain a deviation value of an actual vehicle body and the planned path, and adjusting the running track of the intelligent vehicle according to the running path and the deviation value; the tracking precision of the AGV intelligent vehicle is improved.
Drawings
FIG. 1 is a schematic flowchart of a method for controlling an AGV based on visual guidance according to embodiment 1 of the present invention;
fig. 2 is a schematic structural diagram of a control system for guiding an AGV smart vehicle based on vision according to embodiment 2 of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Example 1
The embodiment of the invention provides a control method for guiding an AGV intelligent vehicle based on vision, which has a flow schematic diagram, and as shown in FIG. 1, the control method for guiding the AGV intelligent vehicle based on vision comprises the following steps:
s1, acquiring a road image of a road where the intelligent vehicle is located, preprocessing the road image to obtain a preprocessed image, and scanning and comparing the preprocessed image to acquire the boundary and the road condition of the road where the intelligent vehicle is located;
step S2, acquiring road center line information according to the boundary of the road where the intelligent vehicle is located and the road condition, and planning the driving path of the intelligent vehicle according to the center line information;
and S3, carrying out weighting and summing processing according to the running path to obtain a deviation value of the actual vehicle body and the planned path, and adjusting the running track of the intelligent vehicle according to the running path and the deviation value.
Preferably, the road image is preprocessed to obtain a preprocessed image, the preprocessed image is scanned and compared to obtain the boundary and the road condition of the road where the intelligent vehicle is located, and the method specifically comprises the following steps,
carrying out graying, binarization and compression processing on the road image, separating the outlines of a road and a non-road to obtain a preprocessed image, scanning the middle of the road outline in the preprocessed image to two sides to obtain pixel jumping points to obtain the boundary of the road where the intelligent vehicle is located, and obtaining the road condition of the intelligent vehicle according to the boundary of the road where the intelligent vehicle is located, wherein the road condition is any one of a straight road, a cross road, a roundabout road and a curve road.
In a specific embodiment, the intelligent vehicle host reads an image transmitted by a camera frame by frame, calls IMREAD in opencv during reading, reads a parameter set IMREAD _ GRAYSCALE, grays the image and stores the image by a mat type variable in the opencv; then, a threshold function is called, CV _ THRESH _ OTSU is selected as a parameter, and an Otsu threshold value is selected to carry out binarization on the image, wherein the Otsu threshold value can remarkably reduce the negative influence of over-strong or over-weak light on the image; finally, calling a resize function, and compressing the initial image to 60 multiplied by 80 pixel size by adopting a nearest neighbor interpolation method, and separating the outlines of the road and the non-road, thereby realizing the preprocessing of the image;
scanning and comparing the preprocessed pictures to obtain the boundary of the road, and according to the boundary of the road, adopting a direct edge detection algorithm to scan the pictures which are binarized line by line from the middle to two sides of the road, wherein the scanned jumping points of pixels are the boundary of the road, setting two one-dimensional arrays of Left 60 and Right 60 for storage respectively, if the jumping points are not scanned, the boundary of the line is lost, setting 0 to mark;
preferably, the method comprises the steps of obtaining road center line information according to the boundary of the road where the intelligent vehicle is located and the road condition, planning the driving path of the intelligent vehicle according to the center line information, and specifically comprising,
step S10, dividing the preset road into n rows and m columns to form an n x m array on the preset road, and pre-storing the width of each row of the preset road in the n x m size graph;
step S20, if the road condition of the intelligent vehicle is a straight road, a cross road or a curve road, executing step S40, and if the road condition of the intelligent vehicle is a rotary island, executing step S30;
step S30, connecting the pixel jumping point with the nearest left boundary point of the image, and taking the connecting line as the left boundary;
and step S40, regarding the boundary of the road, if the left boundary and the right boundary exist, the midpoint of the corresponding line in the road is the average value of the corresponding numerical values of the two boundaries, if the left boundary does not have the right boundary, the left boundary adds half of the road width of the corresponding line to obtain the midpoint of the corresponding line, if the right boundary does not have the left boundary, the right boundary subtracts half of the road width of the corresponding line to obtain the midpoint of the corresponding line, if the left boundary and the right boundary are both boundaries, the midpoint of the previous line is taken as the midpoint of the current line, the midpoints corresponding to all the lines are connected to obtain center line information, and the center line information is the planned driving path of the intelligent vehicle.
In a specific embodiment, due to the particularity of the road condition of the roundabout, a quite obvious abrupt angle exists in an image before entering the roundabout, lines are lost by comparing the number of lines with the number of lines which is larger than 30 and the right boundary of the line during scanning, and the upper two lines are all non-road parts at the boundary, namely the roundabout is judged;
because different road conditions correspond to different driving requirements, different path planning must be carried out on different road conditions, and based on actual conditions, the road conditions have: the number of the straight road, the cross, the rotary island and the bend is four;
the embodiment of the invention provides a method for presetting road width, which firstly stores the width of each road in a 60 x 80 size graph at the beginning of starting an intelligent vehicle, so that the processing of straight roads, crosses and curves can be finished by one method, and the specific operation is as follows:
for the case that the edge is scanned to the boundary, if the left and right boundaries exist, the midpoint of the corresponding row is equal to the average of the corresponding values of the left and right boundaries; if the left side has a boundary and the right side has no boundary, the middle point of the line can be obtained by adding a preset half of the width of the road of the line by the left boundary, if the right side has a boundary and the left side has no boundary, the middle point of the line can be obtained by subtracting a preset half of the width of the road of the line changing by the right boundary, and if the middle points do not exist on the two sides, the middle point of the previous line at the near end in the image is taken as the middle point of the line;
the method can not only reduce the design amount of the algorithm, but also accurately complete the path planning of road conditions such as straight roads, crosses, curved roads, rotary islands and the like without any characteristic judgment;
for the road condition of the rotary island, connecting the scanned characteristic points (pixel jumping points) and the nearest left boundary point of the image (the view field of the camera), taking the connecting line as the left boundary line of the road, and then planning the path by using a path planning method of three road conditions of straight road, cross road and curved road; all planned paths are stored in the center [60] array;
preferably, the weighted summation is performed according to the driving path to obtain the deviation value between the actual vehicle body and the planned path, and specifically, the weighted summation is performed on the center of each row of the preset road, and the difference is made between the summation result and the road position center line to obtain the deviation value between the actual vehicle body and the planned path.
During specific implementation, the intelligent vehicle control steering engine is adjusted according to the deviation value, and real-time adjustment is carried out, so that stable and rapid running of the intelligent vehicle is realized; the central line of the position of the car body, namely the 40 th column of the visual field, is a fixed value, and the position of the central line of the position of a road (track) in a camera is changed from time to time, namely, the difference value between the central line of the predicted road position and the actual central line of the intelligent car is required to be taken, and the steering engine angle control is carried out according to the difference value;
in a specific embodiment, the midpoint of each row of the center [60] is weighted, and the weighting value principle is that for 15 rows at the near end of the visual field, an intelligent vehicle is about to drive in a very short time, the prediction influence on the road position central line (preset road middle column number) is small, the weight value assignment is small, the row 16-45 is interrupted, the distance from the intelligent vehicle is about 1m-1.3m, the distance from the intelligent vehicle to the intelligent vehicle is about to pass in a certain time in the future, the prediction influence on the road position central line position is large, the weight value assignment is large, and for the far end 45-60, the prediction influence on the road position central line is small due to the over-far distance and the far end is easy to generate visual field distortion, and the weight value assignment is minimum; weighting, calculating a weighted sum, and performing difference with 40 to obtain a deviation value between the intelligent vehicle and the road so as to control a steering engine;
preferably, the control method for guiding the AGV smart vehicle based on the vision further includes detecting whether an obstacle exists, if so, acquiring an index position of the obstacle in an n × m array, determining whether the obstacle is on the left side or the right side of the smart vehicle, taking a forward direction of the smart vehicle as a positive direction of an x axis, acquiring a vertical distance from the obstacle to the x axis, and if a ratio of the vertical distance to a half vehicle width of the smart vehicle is smaller than a preset value, enabling the smart vehicle to avoid the obstacle to the right side or the left side according to whether the obstacle is on the left side or the right side of the smart vehicle.
In specific implementation, the index position of the obstacle in an n × m array (specifically, a 60 × 80 array) is obtained, whether the obstacle is on the left side or the right side of the intelligent vehicle is determined, the advancing direction of the intelligent vehicle is the forward direction of an x axis, the vertical distance between the obstacle and the x axis is further found, the danger depth is measured by the ratio of the vertical distance between the intelligent vehicle and the x axis to a half vehicle width, if the ratio of the vertical distance to the half vehicle width of the intelligent vehicle is smaller than a preset value (such as 1), namely, the vertical distance between the obstacle and the x axis is smaller than the half vehicle width of the intelligent vehicle, the fact that the intelligent vehicle is in a straight line collision at the moment is indicated, the obstacle is dangerous, and the intelligent vehicle avoids the obstacle to the right side or the left side according to whether the obstacle is on the left side or the right side of the intelligent.
Example 2
The invention provides a control system for guiding an AGV (automatic guided vehicle) based on vision, which has a schematic structural diagram, and as shown in FIG. 2, the control system for guiding the AGV based on vision comprises a road state acquisition module, a driving path planning module and a driving track adjustment module;
the road state acquisition module is used for acquiring a road image of a road where the intelligent vehicle is located, preprocessing the road image to obtain a preprocessed image, scanning and comparing the preprocessed image to acquire the boundary and the road condition of the road where the intelligent vehicle is located;
the driving path planning module is used for acquiring road center line information according to the boundary of the road where the intelligent vehicle is located and the road condition, and planning the driving path of the intelligent vehicle according to the center line information;
and the running track adjusting module is used for carrying out weighting and summing processing according to the running path, acquiring a deviation value of an actual vehicle body and the planned path, and adjusting the running track of the intelligent vehicle according to the running path and the deviation value.
In specific implementation, the road state acquisition module specifically comprises a camera and an image preprocessing module, wherein the camera is a USB high frame rate camera and is placed at a position 30cm away from the ground from the center of the vehicle; the camera is used for acquiring image information and transmitting the image information to an intelligent vehicle host (an embedded single chip microcomputer or a microprocessor) for image preprocessing; storing an image collected by a camera by using Mat type data (480 × 360 size of 3 channels) in OpenCV, calling OpenCV to carry out preprocessing operations such as interception, binarization, graying, compression and the like on the image, for example, calling cv:: Rect function to intercept the main part of a road, and compressing the intercepted part to 60:.80 size by adopting INTER _ LINEAR bilinear interpolation method of resize function; and then the gray level and the filtering processing of the compressed image are finished through cvtColor and GaussianBlur, and finally the binaryzation of the Dajin threshold value is realized through threshold, so that the influence of nonuniform light on the binaryzation of the image can be reduced. Finally, obtaining a 60 × 80 binary road picture without noise points;
acquiring the boundary and the road condition of a road where the intelligent vehicle is located; the processed image array is transmitted into a driving path planning module in a form of message in ROS so as to carry out center line calculation and road condition identification; image preprocessing is realized in a ROS standard node mode, reading and preprocessing of images are completed by using OpenCV, the images are converted into standard ROS messages, detection of left and right edges is completed on the images, and finally an image array and the left and right edges are borne by a self-defined message and are issued to the overall ROS;
preferably, the road state obtaining module is configured to pre-process the road image to obtain a pre-processed image, scan and compare the pre-processed image to obtain a boundary and a road condition of a road where the intelligent vehicle is located, and specifically includes,
carrying out graying, binarization and compression processing on the road image, separating the outlines of a road and a non-road to obtain a preprocessed image, scanning the middle of the road outline in the preprocessed image to two sides to obtain pixel jumping points to obtain the boundary of the road where the intelligent vehicle is located, and obtaining the road condition of the intelligent vehicle according to the boundary of the road where the intelligent vehicle is located, wherein the road condition is any one of a straight road, a cross road, a roundabout road and a curve road.
Preferably, the driving path planning module acquires road center line information according to the boundary of the road where the intelligent vehicle is located and the road condition, and plans the driving path of the intelligent vehicle according to the center line information, and specifically includes,
step S10, dividing the preset road into n rows and m columns to form an n x m array on the preset road, and pre-storing the width of each row of the preset road in the n x m size graph;
step S20, if the road condition of the intelligent vehicle is a straight road, a cross road or a curve road, executing step S40, and if the road condition of the intelligent vehicle is a rotary island, executing step S30;
step S30, connecting the pixel jumping point with the nearest left boundary point of the image, and taking the connecting line as the left boundary;
and step S40, regarding the boundary of the road, if the left boundary and the right boundary exist, the midpoint of the corresponding line in the road is the average value of the corresponding numerical values of the two boundaries, if the left boundary does not have the right boundary, the left boundary adds half of the road width of the corresponding line to obtain the midpoint of the corresponding line, if the right boundary does not have the left boundary, the right boundary subtracts half of the road width of the corresponding line to obtain the midpoint of the corresponding line, if the left boundary and the right boundary are both boundaries, the midpoint of the previous line is taken as the midpoint of the current line, the midpoints corresponding to all the lines are connected to obtain center line information, and the center line information is the planned driving path of the intelligent vehicle.
Preferably, the driving track adjusting module performs weighted summation processing according to the driving path to obtain the deviation value between the actual vehicle body and the planned path, and specifically includes performing amplitude weighting processing and weighted summation on the center of each row of the preset road, and performing difference between the summation result and the road position center line to obtain the deviation value between the actual vehicle body and the planned path.
Preferably, the control system for guiding the AGV intelligent vehicle based on the vision further comprises an obstacle avoidance module, wherein the obstacle avoidance module is used for detecting whether an obstacle exists or not, if the obstacle exists, the index position of the obstacle in an n x m array is obtained, the obstacle is determined to be on the left side or the right side of the intelligent vehicle, the advancing direction of the intelligent vehicle is the positive direction of the x axis, the vertical distance from the obstacle to the x axis is obtained, and if the ratio of the vertical distance to the half vehicle width of the intelligent vehicle is smaller than a preset value, the intelligent vehicle is enabled to avoid the obstacle to the right side or the left side according to the fact that the obstacle exists on the left side or the right side of the intelligent vehicle.
In specific implementation, a laser radar is used for detecting whether obstacles exist or not, in order to ensure that information such as the obstacles does not influence the identification of road conditions, the laser radar transmits point cloud data of the radar to a data processing unit in an obstacle avoidance module, the data processing unit in the obstacle avoidance module calculates the direction and the property of the current obstacle, and finally information is transmitted to a road state acquisition module, so that the road condition information cannot be misjudged under the condition that the obstacles exist; the laser radar is driven by a standard driving package of the ROS, the scanning frequency is 10Hz, two-dimensional point cloud data are obtained, data in the range of 60 degrees in the forward direction and the reverse direction of the radar are taken, and messages are written into a standard array form in the ROS to be issued;
a data processing unit in the obstacle avoidance module acquires information of the laser radar, wherein the information comprises obstacle points in a range of 360 degrees of a scanning plane of the laser radar, the scanning step length is 0.5 degree, namely, a one-dimensional array with the length of 720 is acquired every time the laser radar is scanned, and the numerical value of each element is the distance of the obstacle points;
because only the obstacles in a certain range in the front-moving direction of the intelligent vehicle can influence the intelligent vehicle, the range of 60 degrees before and after the intelligent vehicle is taken for data processing, namely, the one-dimensional arrays with the use length of 120 before and after the intelligent vehicle are used for storing partial data of the laser radar;
solving the surface curvature of the obstacle by utilizing a leastsq function of python, taking an array in a forward direction as an example, detecting the number of obstacle points within 2m from 120 obstacle points in the 60-degree range, determining that the intelligent vehicle meets the threatened obstacle once the value exceeds a threshold value, converting the absolute distance of the obstacle points into coordinate values of x and y axes in a laser radar coordinate system, giving the angle of each obstacle point through the corresponding relation between the index number and the angle of the array, and transferring the coordinate values of the points to the leastsq () function to solve the fitted curvature;
the center distance of the obstacle is directly read in an array, the width of the obstacle is calculated by converting the center distance between a starting obstacle point and an ending obstacle point into coordinate values of x and y, and the absolute value of the x-direction coordinate is the width of the obstacle. The calculation of the danger degree is that firstly, the index position of an obstacle point in an array is checked, whether the obstacle point is on the left side or the right side of the intelligent vehicle is determined, the advancing direction of the intelligent vehicle is the x forward direction, the obstacle point with the minimum vertical distance of the x axis is further found, if the distance from the point to the x axis is smaller than half of the vehicle width of the intelligent vehicle, the fact that the intelligent vehicle is in collision when the intelligent vehicle is in straight running is indicated, the obstacle is dangerous, and the danger degree is measured by the ratio of the vertical distance of the obstacle point to the x axis to the half of the vehicle width;
in specific implementation, the image preprocessing module, the path planning module, the driving track adjusting module and the obstacle avoidance module can be integrated in a second microprocessor (i.mx6 model), the control system for guiding the AGV intelligent vehicle based on the vision also comprises a speed calculating module, a communication module and a power supply module, the speed calculating module can be arranged in the second microprocessor, the speed calculating module acquires information such as a driving path, a deviation value and whether to avoid an obstacle or not, calculates an expected speed and an expected rotation angle, and transmits data to the first microprocessor (MPC5744p) in real time through the communication module; the direct current motor driving module, the steering module, the speed detection module, the power supply module and the communication module are connected to the first microprocessor; the first microprocessor sends a control instruction through operation, and controls the speed and the rotation angle of the vehicle body through the direct current motor driving module, the steering module and the speed detection module, so that tracking and obstacle avoidance are realized; the barrier recognition and the planning of various paths are realized through the fusion of the camera and the laser radar, and the intelligent vehicle based on the visual guidance is completed
The control system for guiding the AGV intelligent vehicle based on the vision also comprises a direct current motor driving module, a steering module, a speed detection module and a power supply module; the direct current motor driving module is connected to direct current motors on two sides of the tail of the vehicle body to drive rear wheels of the vehicle body to move, the steering module is a steering engine steering module and is connected to wheels on the front portion of the vehicle body to drive the vehicle body to steer, the speed detection module is an encoder speed detection module and is connected to wheels on two sides of the rear portion of the vehicle body, the direct current motor driving module, the steering module and the speed detection module are connected to a first microprocessor, and the communication module is a serial port communication module and is connected to the first microprocessor and a second microprocessor to achieve communication; the final speed information of the intelligent vehicle is sent to the bottom layer single chip microcomputer through the communication module; through first microprocessor and second microprocessor are connected to different modules, full play the treater performance, also can improve information processing speed, have increased intelligent car and have sought mark precision and interference killing feature
In a specific embodiment, not only is the coordinate information of the obstacle sent by the obstacle avoidance module, but also a function of solving a homography matrix of the camera by using OpenCV is used, so that the coordinate of an actual coordinate plane is solved by a pixel coordinate plane by using a monocular camera, and the coordinate plane is consistent with a plane where the laser radar is located; the detection accuracy is improved by the fusion of the laser radar and the camera when the obstacle is detected again;
the direct current motor driving module comprises a left driving motor and a right driving motor at the rear part of the vehicle body, the motor model is an RN-380 direct current motor, the rotating speed is controlled by PWM pulse width modulation, and the PWM frequency is 10 KHZ; the first microprocessor receives an expected speed control instruction of the second microprocessor through the communication module, and a timer module in the first microprocessor outputs a digital signal to control the armature voltage of the motor, so that the aim of controlling the rotating speed is fulfilled;
the steering module adopts an FUTABA3010 servo motor as a steering engine to control the steering of the vehicle body, angle control is carried out by means of PWM pulse width modulation, and the PWM frequency is 50 HZ; the first microprocessor receives an expected angle control instruction of the second microprocessor through the communication module, and a timer module in the first microprocessor outputs a digital signal to control the armature voltage of the motor, so that the aim of angle control is fulfilled;
the speed detection module Mini1024Z biphase incremental encoder; and the pulse capture of the encoder is realized by means of a timer module in the first microprocessor, so that the aim of speed detection is fulfilled.
The communication module adopts a CH340G chip, the communication protocol adopts universal asynchronous Receiver/Transmitter (UART), and the high-speed stable communication between the first microprocessor and the second microprocessor is achieved by setting relevant parameters such as baud rate, data bit, stop bit, parity bit and the like; the speed control strategy in the direct current motor driving module adopts an incremental PID control strategy; the incremental PID control strategy takes the difference value between the actual speed acquired by the Mini1024Z two-phase incremental encoder and the speed transmitted by the second microprocessor through the communication module as input, takes the digital signal PWM pulse width output by the timer module in the first sensor as output, and controls the armature voltage of the motor through the direct current motor driving module to achieve the purpose of quickly achieving the desired speed.
The invention discloses a method and a system for controlling an AGV (automatic guided vehicle) based on visual guidance, wherein the method comprises the steps of acquiring a road image of a road where the AGV is located, preprocessing the road image to obtain a preprocessed image, scanning and comparing the preprocessed image to acquire the boundary and the road condition of the road where the AGV is located; acquiring road center line information according to the boundary of the road where the intelligent vehicle is located and the road condition, and planning a driving path of the intelligent vehicle according to the center line information; carrying out weighted summation processing according to the running path to obtain a deviation value of an actual vehicle body and the planned path, and adjusting the running track of the intelligent vehicle according to the running path and the deviation value; according to the scheme, the design quantity of the algorithm is reduced, meanwhile, path planning of different road conditions can be accurately finished, the tracking precision of the AGV intelligent vehicle is improved, the anti-interference capacity of the AGV intelligent vehicle is improved, and the obstacle avoidance capacity of the AGV intelligent vehicle is improved by detecting obstacles;
the above-described embodiments of the present invention should not be construed as limiting the scope of the present invention. Any other corresponding changes and modifications made according to the technical idea of the present invention should be included in the protection scope of the claims of the present invention.
Claims (10)
1. A control method for guiding an AGV intelligent vehicle based on vision is characterized by comprising the following steps:
acquiring a road image of a road where an intelligent vehicle is located, preprocessing the road image to obtain a preprocessed image, scanning and comparing the preprocessed image to acquire a boundary and a road condition of the road where the intelligent vehicle is located;
acquiring road center line information according to the boundary of the road where the intelligent vehicle is located and the road condition, and planning a driving path of the intelligent vehicle according to the center line information;
and carrying out weighted summation processing according to the running path to obtain a deviation value of the actual vehicle body and the planned path, and adjusting the running track of the intelligent vehicle according to the running path and the deviation value.
2. The method for controlling the AGV according to claim 1, wherein the images of the road are pre-processed to obtain pre-processed images, the pre-processed images are scanned and compared to obtain boundaries and road conditions of the road where the intelligent vehicle is located, and the method specifically comprises,
carrying out graying, binarization and compression processing on the road image, separating the outlines of a road and a non-road to obtain a preprocessed image, scanning the middle of the road outline in the preprocessed image to two sides to obtain pixel jumping points to obtain the boundary of the road where the intelligent vehicle is located, and obtaining the road condition of the intelligent vehicle according to the boundary of the road where the intelligent vehicle is located, wherein the road condition is any one of a straight road, a cross road, a roundabout road and a curve road.
3. The method for controlling the AGV according to claim 2, wherein the method comprises the steps of obtaining road center line information according to the boundary of the road where the intelligent vehicle is located and the road condition, and planning the driving path of the intelligent vehicle according to the center line information,
dividing the preset road into n rows and m columns to form an n x m array of the preset road, and pre-storing the width of each row of the preset road in an n x m size graph;
if the road condition of the intelligent vehicle is a straight road, a cross road or a curve road, then
For the boundary of the road, if the left boundary and the right boundary exist, the midpoint of the corresponding line in the road is the average value of the corresponding numerical values of the two boundaries, if the left boundary does not have the right boundary, the midpoint of the corresponding line is obtained by adding half of the width of the road of the corresponding line to the left boundary, if the right boundary does not have the left boundary, the midpoint of the corresponding line is obtained by subtracting half of the width of the road of the corresponding line from the right boundary, if the left boundary and the right boundary are both boundaries, the midpoint of the previous line is taken as the midpoint of the current line, the corresponding midpoints of all the lines are connected to obtain centerline information, the centerline information is the planned driving path of the intelligent vehicle,
if the road condition of the intelligent vehicle is the roundabout, then
And connecting the pixel jumping points with the left boundary point at the most proximal end of the image, taking the connecting line as a left boundary, and then planning the driving path of the intelligent vehicle when the road condition of the intelligent vehicle is a straight road, a cross road or a curve.
4. The method for controlling the AGV according to claim 3, wherein the step of performing weighted summation processing on the driving path to obtain the deviation value between the actual vehicle body and the planned path specifically comprises the step of performing amplitude value processing and weighted summation on the center of each row of a preset road, and performing difference between the summation result and a road position center line to obtain the deviation value between the actual vehicle body and the planned path.
5. The method as claimed in claim 3, further comprising detecting whether an obstacle exists, if so, obtaining an index position of the obstacle in the n x m array, determining whether the obstacle is on the left side or the right side of the smart vehicle, taking a forward direction of the smart vehicle as a positive direction of an x-axis, obtaining a vertical distance from the obstacle to the x-axis, and if a ratio of the vertical distance to a half vehicle width of the smart vehicle is smaller than a preset value, enabling the smart vehicle to avoid the obstacle to the right side or the left side according to whether the obstacle is on the left side or the right side of the smart vehicle.
6. A control system for guiding an AGV intelligent vehicle based on vision is characterized by comprising a road state acquisition module, a driving path planning module and a driving track adjusting module;
the road state acquisition module is used for acquiring a road image of a road where the intelligent vehicle is located, preprocessing the road image to obtain a preprocessed image, scanning and comparing the preprocessed image to acquire the boundary and the road condition of the road where the intelligent vehicle is located;
the driving path planning module is used for acquiring road center line information according to the boundary of the road where the intelligent vehicle is located and the road condition, and planning the driving path of the intelligent vehicle according to the center line information;
and the running track adjusting module is used for carrying out weighting and summing processing according to the running path, acquiring a deviation value of an actual vehicle body and the planned path, and adjusting the running track of the intelligent vehicle according to the running path and the deviation value.
7. The system of claim 1, wherein the road status acquisition module is configured to pre-process the road image to obtain a pre-processed image, scan and compare the pre-processed image to obtain the boundary and road condition of the road where the intelligent vehicle is located, and specifically comprises,
carrying out graying, binarization and compression processing on the road image, separating the outlines of a road and a non-road to obtain a preprocessed image, scanning the middle of the road outline in the preprocessed image to two sides to obtain pixel jumping points to obtain the boundary of the road where the intelligent vehicle is located, and obtaining the road condition of the intelligent vehicle according to the boundary of the road where the intelligent vehicle is located, wherein the road condition is any one of a straight road, a cross road, a roundabout road and a curve road.
8. The system of claim 7, wherein the driving path planning module obtains road center line information according to the boundary of the road where the intelligent vehicle is located and the road condition, and plans the driving path of the intelligent vehicle according to the center line information, and specifically comprises,
dividing the preset road into n rows and m columns to form an n x m array of the preset road, and pre-storing the width of each row of the preset road in an n x m size graph;
if the road condition of the intelligent vehicle is a straight road, a cross road or a curve road, then
For the boundary of the road, if the left boundary and the right boundary exist, the midpoint of the corresponding line in the road is the average value of the corresponding numerical values of the two boundaries, if the left boundary does not have the right boundary, the midpoint of the corresponding line is obtained by adding half of the width of the road of the corresponding line to the left boundary, if the right boundary does not have the left boundary, the midpoint of the corresponding line is obtained by subtracting half of the width of the road of the corresponding line from the right boundary, if the left boundary and the right boundary are both boundaries, the midpoint of the previous line is taken as the midpoint of the current line, the corresponding midpoints of all the lines are connected to obtain centerline information, and the centerline information is the planned driving path of the intelligent vehicle
If the road condition of the intelligent vehicle is the roundabout, then
And connecting the pixel jumping points with the left boundary point at the most proximal end of the image, taking the connecting line as a left boundary, and then planning the driving path of the intelligent vehicle when the road condition of the intelligent vehicle is a straight road, a cross road or a curve.
9. The system of claim 8, wherein the driving track adjusting module performs weighted summation according to the driving path to obtain the deviation between the actual vehicle body and the planned path, and specifically comprises performing amplitude weighting and weighted summation on the center of each row of the preset road, and performing difference between the summation result and the road position center line to obtain the deviation between the actual vehicle body and the planned path.
10. The control system of claim 8, further comprising an obstacle avoidance module, wherein the obstacle avoidance module is configured to detect whether an obstacle exists, if the obstacle exists, obtain an index position of the obstacle in the n x m array, determine whether the obstacle is on the left side or the right side of the smart vehicle, obtain a vertical distance from the obstacle to the x axis with a forward direction of the smart vehicle as a positive x-axis direction, and if a ratio of the vertical distance to a half vehicle width of the smart vehicle is smaller than a preset value, enable the smart vehicle to avoid the obstacle to the right side or the left side according to whether the obstacle is on the left side or the right side of the smart vehicle.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010475734.0A CN111731324A (en) | 2020-05-29 | 2020-05-29 | Control method and system for guiding AGV intelligent vehicle based on vision |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010475734.0A CN111731324A (en) | 2020-05-29 | 2020-05-29 | Control method and system for guiding AGV intelligent vehicle based on vision |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111731324A true CN111731324A (en) | 2020-10-02 |
Family
ID=72646531
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010475734.0A Pending CN111731324A (en) | 2020-05-29 | 2020-05-29 | Control method and system for guiding AGV intelligent vehicle based on vision |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111731324A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2023201954A1 (en) * | 2022-04-21 | 2023-10-26 | 合众新能源汽车股份有限公司 | Roundabout path planning method and apparatus |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104166400A (en) * | 2014-07-11 | 2014-11-26 | 杭州精久科技有限公司 | Multi-sensor fusion-based visual navigation AGV system |
| CN105046198A (en) * | 2015-06-12 | 2015-11-11 | 上海修源网络科技有限公司 | Lane detection method |
| US20170345190A1 (en) * | 2016-05-27 | 2017-11-30 | Toshiba Medical Systems Corporation | Computed tomography apparatus and empirical pre-weighting method for decreasing image noise nonuniformity |
| CN107505946A (en) * | 2017-10-11 | 2017-12-22 | 安徽建筑大学 | Intelligent carriage path identifying system based on black and white camera |
| CN109033932A (en) * | 2018-05-23 | 2018-12-18 | 华南师范大学 | A kind of racing track recognition methods, identifying system, intelligent vehicle patrol mark method and patrol mark system |
| CN109740532A (en) * | 2018-12-29 | 2019-05-10 | 河海大学常州校区 | A kind of Path Recognition and middle line optimization method based on annulus road |
| CN110132288A (en) * | 2019-05-08 | 2019-08-16 | 南京信息工程大学 | A kind of minicar vision navigation method on wide road surface |
-
2020
- 2020-05-29 CN CN202010475734.0A patent/CN111731324A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104166400A (en) * | 2014-07-11 | 2014-11-26 | 杭州精久科技有限公司 | Multi-sensor fusion-based visual navigation AGV system |
| CN105046198A (en) * | 2015-06-12 | 2015-11-11 | 上海修源网络科技有限公司 | Lane detection method |
| US20170345190A1 (en) * | 2016-05-27 | 2017-11-30 | Toshiba Medical Systems Corporation | Computed tomography apparatus and empirical pre-weighting method for decreasing image noise nonuniformity |
| CN107505946A (en) * | 2017-10-11 | 2017-12-22 | 安徽建筑大学 | Intelligent carriage path identifying system based on black and white camera |
| CN109033932A (en) * | 2018-05-23 | 2018-12-18 | 华南师范大学 | A kind of racing track recognition methods, identifying system, intelligent vehicle patrol mark method and patrol mark system |
| CN109740532A (en) * | 2018-12-29 | 2019-05-10 | 河海大学常州校区 | A kind of Path Recognition and middle line optimization method based on annulus road |
| CN110132288A (en) * | 2019-05-08 | 2019-08-16 | 南京信息工程大学 | A kind of minicar vision navigation method on wide road surface |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2023201954A1 (en) * | 2022-04-21 | 2023-10-26 | 合众新能源汽车股份有限公司 | Roundabout path planning method and apparatus |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP3418943B1 (en) | Object detecting apparatus, object detecting method, and computer-readable medium | |
| CN110243372B (en) | Intelligent agricultural machinery navigation system and method based on machine vision | |
| EP1193661B1 (en) | Lane recognition apparatus for vehicle | |
| US8736820B2 (en) | Apparatus and method for distinguishing ground and obstacles for autonomous mobile vehicle | |
| CN110054116B (en) | Fork navigation method and system applied to forklift and unmanned forklift | |
| CN110745140B (en) | Vehicle lane change early warning method based on continuous image constraint pose estimation | |
| CN103891697B (en) | The variable spray method of a kind of indoor autonomous spraying machine device people | |
| CN112537294B (en) | Automatic parking control method and electronic equipment | |
| CN106313046A (en) | Multi-level obstacle avoidance system of mobile robot | |
| JP6868699B2 (en) | Roadside object detection device, roadside object detection method and roadside object detection system | |
| CN109033932B (en) | Track identification method, track identification system, intelligent vehicle track patrol method and track patrol system | |
| CN111731324A (en) | Control method and system for guiding AGV intelligent vehicle based on vision | |
| Takahashi et al. | Rear view lane detection by wide angle camera | |
| CN113258638A (en) | Intelligent robot charging method and system | |
| CN115562281A (en) | Dynamic path planning method for automatic parking system in complex environment | |
| CN113805571B (en) | Robot walking control method, system, robot and readable storage medium | |
| CN113268065B (en) | AGV self-adaptive turning obstacle avoidance method, device and equipment based on artificial intelligence | |
| JP5461494B2 (en) | Automated traveling vehicle and control method for automated traveling vehicle | |
| CN218536835U (en) | Automatic parking and parking space identification system | |
| CN114661051A (en) | Front obstacle avoidance system based on RGB-D | |
| Aggarwal et al. | Vision based collision avoidance by plotting a virtual obstacle on depth map | |
| WO2021051736A1 (en) | Method and apparatus for determining sensing area, and storage medium and vehicle | |
| 後方カメラ用画像処理技術 et al. | Image processing technology for rear view camera (1): Development of lane detection system | |
| JP3585948B2 (en) | Travel control method for autonomous traveling work vehicle | |
| CN112306064A (en) | RGV control system and method for binocular vision identification |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20201002 |