CN113076910B - AGV trolley magnetic tape interferent detection method and system based on artificial intelligence - Google Patents
AGV trolley magnetic tape interferent detection method and system based on artificial intelligence Download PDFInfo
- Publication number
- CN113076910B CN113076910B CN202110412844.7A CN202110412844A CN113076910B CN 113076910 B CN113076910 B CN 113076910B CN 202110412844 A CN202110412844 A CN 202110412844A CN 113076910 B CN113076910 B CN 113076910B
- Authority
- CN
- China
- Prior art keywords
- interferent
- trolley
- information
- sensor
- interference
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
Abstract
The invention relates to the technical field of artificial intelligence, in particular to an AGV trolley magnetic tape interferent detection method and system based on artificial intelligence. The method comprises the steps of collecting images under a front-view path of a trolley through a camera to obtain RGB images and depth images; obtaining interferent information by analyzing the RGB image; analyzing the interferent information by using a pre-trained offset prediction network to obtain an offset angle, and correcting the position of the trolley according to the offset angle; and predicting the predicted distance information of each sampling point of the sensor and the interferent after the trolley is corrected, and performing weighting adjustment on the sensitivity of the sampling point of the sensor according to the distance information and the magnetic induction intensity information received by the sampling point, so that the trolley can work stably. The invention reduces the influence of the interferent on the operation of the trolley through image processing and sensor weighting adjustment, so that the trolley can stably complete the work task.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to an AGV trolley magnetic tape interferent detection method and system based on artificial intelligence.
Background
Along with the degree of automation is higher and higher, the application field of automatic navigation carrier is also more extensive, and its salient characteristic is automatic navigation. There are many kinds of navigation modes of AGVs, and tape guidance is still widely used in many fields due to its advantages of low cost, simple laying, high navigation precision, easy path change and addition, etc.
The magnetic tape guide transport vehicle collects magnetic tape signals through a magnetic navigation sensor on the vehicle body and runs along the magnetic tape laid on the ground. Aiming at the storage system of some metal products or devices, metal can generate magnetism on the magnetic guide leading belt in the transportation process so as to cause interference on the magnetic guide leading belt, and if interference information is not checked in time, the trolley can yaw to cause a series of consequences. Therefore, the problem to be solved at present is to detect the information of the interferent in time and adjust the trolley according to the information of the interferent.
Disclosure of Invention
In order to solve the technical problems, the invention aims to provide an AGV trolley magnetic tape interferent detection method and system based on artificial intelligence, and the adopted technical scheme is as follows:
the invention provides an AGV trolley magnetic tape interferent detection method based on artificial intelligence, which comprises the following steps:
when the magnetic induction intensity signal acquired by the sensor is abnormal, acquiring an RGB image and a depth image under an orthographic path through an RGBD (red, green and blue) camera carried by the trolley;
dividing the RGB image into magnetic tape regions as regions of interest, and performing interferent analysis on the RGB image and the depth image in the regions of interest to obtain interferent images; obtaining interference object information according to the interference object image; the interferent information comprises the area of the interferent, the distance from the interferent to the sensor, the magnetization degree of the interferent and an interference mapping range;
processing the interferent information by using a pre-trained offset prediction network to obtain an offset angle; correcting the position of the trolley by using the offset angle;
predicting the predicted distance information of each sampling point of the sensor and the interferent after the trolley is corrected, wherein the predicted distance information is combined with the magnetic induction intensity information received by the sampling points to perform weighting adjustment on the sensitivity of the sampling points.
Further, the performing interferent analysis on the RGB image and the depth image within the region of interest comprises:
extracting the color, luster and texture of the interferent in the region of interest as interferent characteristics to judge whether the interferent is a metal interferent; if the interferent is a non-metal interferent, recording the current position information of the trolley; and if the interference object is a metal interference object, recording the current position information and simultaneously obtaining the image of the interference object.
Further, the method for obtaining the interference mapping range includes:
and obtaining the length of the interfering object according to the coordinate information in the interfering object image, and taking the ratio of the length of the sensor to the length of the interfering object as the interference mapping range.
Further, the magnetic induction intensity information acquisition method comprises the following steps:
obtaining a plurality of groups of signal clusters by clustering and analyzing the signal value of each sampling point; and calculating the ratio of each cluster of signal values in the total signal value to obtain a signal ratio as the magnetic induction intensity information.
Further, when the sensor collects a magnetic induction intensity signal, if the magnetic induction intensity signal is smaller than a preset magnetic induction intensity threshold value, the trolley is determined not to be on the magnetic belt, and the position of the trolley is corrected.
The invention also provides an AGV trolley magnetic tape interferent detection system based on artificial intelligence, which comprises: the device comprises an image acquisition module, an image analysis module, an offset angle acquisition module and a sensor adjustment module;
the image acquisition module is used for acquiring RGB images and depth images under an orthographic path through an RGBD (red, green and blue) camera carried by the trolley when magnetic induction intensity signals acquired by the sensor are abnormal;
the image analysis module is used for dividing the RGB image into magnetic tape regions as regions of interest, and performing interference object analysis on the RGB image and the depth image in the regions of interest to obtain interference object images; obtaining interference object information according to the interference object image; the interferent information comprises the area of the interferent, the distance from the interferent to the sensor, the magnetization degree of the interferent and an interference mapping range;
the offset angle acquisition module is used for processing the interferent information by utilizing a pre-trained offset prediction network to acquire an offset angle; correcting the position of the trolley by using the offset angle;
the sensor adjusting module is used for predicting the predicted distance information of each sampling point of the sensor and the interferent after the trolley is corrected, and the predicted distance information is combined with the magnetic induction intensity information received by the sampling points to perform weighting adjustment on the sensitivity of the sampling points.
Further, the image analysis module further comprises a semantic segmentation module;
the semantic segmentation module is used for extracting the color, luster and texture of the interferent in the region of interest as interferent characteristics to judge whether the interferent is a metal interferent; if the interferent is a non-metal interferent, recording the current position information of the trolley; and if the interference object is a metal interference object, recording the current position information and simultaneously obtaining the image of the interference object.
Further, the image analysis module further comprises an interference mapping range acquisition module;
the interference mapping range acquisition module is used for acquiring the length of the interference object according to the coordinate information in the interference object image, and the ratio of the length of the sensor to the length of the interference object is used as the interference mapping range.
Further, the sensor adjusting module further comprises a cluster analyzing module;
the cluster analysis module is used for obtaining a plurality of groups of signal clusters by cluster analysis of the signal value of each sampling point; and calculating the ratio of each cluster of signal values in the total signal value to obtain a signal ratio as the magnetic induction intensity information.
Further, the system also includes a tape position correction module;
the magnetic tape position correction module is used for determining that the trolley is not on the magnetic tape and correcting the position of the trolley if the magnetic induction intensity signal is smaller than a preset magnetic induction intensity threshold value when the sensor collects the magnetic induction intensity signal.
The invention has the following beneficial effects:
1. according to the embodiment of the invention, the trolley offset angle is obtained by analyzing the information of the interferent through the offset prediction network, and the position and the advancing of the trolley are corrected according to the trolley offset angle, so that serious errors generated in the working process of the trolley are avoided.
2. According to the embodiment of the invention, the sensors are dynamically weighted through the offset angle, the interferent information and the sensor information, and the sensitivity of the sensors is adjusted, so that the trolley can stably work aiming at interference in the advancing process.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions and advantages of the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
FIG. 1 is a flowchart of an AGV magnetic tape interferent detection method based on artificial intelligence according to an embodiment of the present invention;
FIG. 2 is a block diagram of an AGV car magnetic tape jammer detection system based on artificial intelligence according to an embodiment of the present invention.
Detailed Description
To further illustrate the technical means and effects of the present invention adopted to achieve the predetermined objects, the following detailed description will be given to a method and a system for detecting AGV magnetic tape interferent based on artificial intelligence according to the present invention, with reference to the accompanying drawings and preferred embodiments, and the specific implementation, structure, features and effects thereof. In the following description, the different references to "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The following describes a specific scheme of an AGV car magnetic tape interferent detection method and system based on artificial intelligence in detail with reference to the accompanying drawings.
Referring to fig. 1, a flowchart of an AGV car magnetic tape jammer detection method based on artificial intelligence according to an embodiment of the present invention is shown, where the method includes:
step S1: when the sensor acquires abnormal signals, the RGB image and the depth image under the normal vision path are acquired.
The magnetic induction sensor on the AGV trolley is a sensor comprising a plurality of sampling points, and the sensors are generally arranged at 16 sampling points and are arranged at intervals of 10 mm. When the signals collected by the sensors are abnormal, the AGV trolley carries the RGBD camera to collect images under the current path, and RGB images and depth images containing depth information are obtained. The deployment position of the camera is a front view angle, the view field range can cover the road condition in front of the trolley, and the information of the front tape under the front view angle of the trolley can be obtained.
Preferably, the magnetic induction intensity is detected through the sensor, when the magnetic induction intensity acquired by the sensor suddenly decreases to a preset magnetic induction intensity threshold value, the current trolley traveling position is not in the center of the magnetic tape, the differential steering can be realized by driving the driving force through the servo driving motor, the trolley and the sensor on the trolley are ensured to be in the center of the magnetic tape, and the factor of the magnetic induction intensity change caused by the fact that the trolley is not in the magnetic tape is eliminated in subsequent analysis.
In the embodiment of the present invention, the threshold magnetic induction is set to 80.
Step S2: and analyzing and processing the RGB image to obtain interferent information.
In order to obtain accurate interfering object information, a magnetic tape region in an RGB image is first used as a region of interest. The region of interest can be divided by color difference, or the neural network can be trained to extract and divide the region of interest for the characteristics of the tape region, which is not limited herein.
And after the region of interest is obtained, carrying out interference object analysis on the RGB image and the depth image in the region of interest to obtain interference object information. The interferent information comprises the area of the interferent, the distance from the interferent to the sensor, the degree of magnetization of the interferent, and the interference mapping range.
Preferably, the color, luster and texture of the interferent in the region of interest are extracted as the characteristics of the interferent, and whether the interferent is a metal interferent is judged. If the sensor is a non-metal interference object, the current sensor abnormal signal is caused by the fact that the current magnetic band region is worn and aged, and only the current trolley position is recorded, so that subsequent workers can clean and maintain conveniently. If the interference object is a metal interference object, the position is recorded and an interference object image is output at the same time.
In the embodiment of the invention, the images in the region of interest are analyzed by using a pre-trained semantic segmentation network to obtain the interference object image. The specific training steps of the semantic segmentation network are as follows:
1) A plurality of images containing metallic interferents are acquired as training data. The metal interferents within the training data set the labels. And normalizing the label data and the training data, eliminating the influence brought by singular sample data, and accelerating the speed of solving the optimal solution by gradient descent.
2) And (4) with the color, the luster and the texture as characteristics, sending the training data and the label data after the normalization processing into a network. The network employs an encoding-decoding architecture. The semantic segmentation encoder performs feature extraction on input data, and comprises convolution and downsampling operations, feature mapping is continuously shrunk, and the size is reduced. And then restoring the feature size to the size of the input image through upsampling and jumping connection by a semantic segmentation decoder, increasing feature dimensions through upsampling operation, converting the feature vector into a category label by using 1-by-1 convolution, and judging the nature of the interferent according to the category label.
3) The network loss function adopts a cross entropy loss function.
The specific method for acquiring the interferent information comprises the following steps:
and carrying out binarization processing on the interference object image to obtain an interference object binary image. In the embodiment of the invention, the pixel value of the pixel point representing the interference object is 1. And counting the number of the pixel points with the pixel value of 1 as the area of the interference object.
And obtaining the distance from each sampling point of the sensor to the interference object through the depth information in the depth image, and taking the average value of the distances as the distance from the interference object to the sensor.
And establishing a coordinate system by taking the upper left of the interference object image as the origin of the coordinate system. And taking the difference value of the upper left abscissa and the lower right abscissa of the interference object as the length of the interference object. The length of the sensor is a fixed value. In the process of trolley traveling, the distance between the trolley and the interferent is closer and closer, and the length of the interferent in the image is longer and longer, so that the ratio of the length of the interferent to the length of the sensor can be used as an interference mapping range.
The degree of interferent magnetization is related to the metal properties and temperature. In the embodiment of the invention, the warehousing system is in a constant temperature environment, and the magnetization degree of the interferent can be obtained after the interferent is detected to be metal.
And step S3: and processing the interferent information by using an offset prediction network to obtain an offset angle and correcting the position of the trolley.
In the embodiment of the invention, a time domain convolution network (TCN) is selected as an offset prediction network to process the information of the interferent, so as to obtain the offset angle.
The training method of the offset prediction network comprises the following steps:
1) The camera sampling rate was set to 30 frames per second for 5 consecutive seconds of historical data as a set of training data sets. The first 4 seconds of data were used as a training data and the 5 th second of yaw angle was used as a label data.
2) Setting an input tensor shape of an offset prediction network as a function of interferer information
Wherein B is the number of the selected samples, 4 is four input characteristics which are respectively the distance from the sensor to the metal interferent, the orthographic area of the metal interferent, the magnetization degree of the metal interferent and the action range of the metal interferent on the sensor, and the final output is the yaw angle when the trolley passes through and has the shape of
。
3) The network is trained using a regression-type loss function. A variance loss function is employed in embodiments of the present invention.
The offset radius of the trolley can be obtained according to the rotating speed between the left wheel and the right wheel of the trolley and the distance between the steering wheels of the left wheel and the steering wheels of the right wheel in a period of time:
wherein the content of the first and second substances,
in order to offset the radius of the light beam,
for the speed of the left wheel,
the speed of the right wheel is the speed of the right wheel,
the distance between the left wheel and the right wheel of the trolley.
The offset of the trolley in the transverse direction and the longitudinal direction can be obtained according to the offset radius and the offset angle:
wherein the content of the first and second substances,
the offset of the trolley in the transverse direction is,
is the offset of the trolley in the longitudinal direction,
in order to offset the radius of the beam,
is an offset angle.
And performing offset compensation on the current position of the trolley through the offset, and correcting the trolley to the center of the route.
And step S4: and carrying out weighting adjustment on the sensitivity of the sampling point.
In the moving process of the trolley, because the yaw angle changes nonlinearly, the signal value received by the sensor changes all the time, and therefore, a weight needs to be dynamically allocated to the sensitivity of each sampling point of the sensor, so that the trolley can stably move.
Preferably, the present invention adopts a clustering method to cluster the sampling signals at each time, and the clustering step is as follows:
1) Collecting the signal value of each sampling point, randomly selecting a signal value as an initial clustering center point
。
2) Calculating the distance between each residual signal value and the initial cluster central point
。
3) At the maximum distance
As new cluster center
. According to the real-time received signal intensity of the sensor in the running process of the trolley
Setting a signal threshold
. When the temperature is higher than the set temperature
And when the selection is finished, the central points of all the clusters are shown to be selected. In an embodiment of the invention, the signal threshold value
And taking 5.
4) The operations 2,3 are repeated until all cluster center points are selected. And finally, obtaining I groups of signal clusters according to the convergence result, wherein the difference value of signal points in each cluster is not more than 5.
5) Calculating the total value of each cluster signal value
Then, the ratio of each cluster of signal values in the total signal value is:
wherein the content of the first and second substances,
for the total value of the signal values of each cluster,
is a first
The signal values of the individual sampling points are,
the number of sample points.
The signal value ratio is used as sensor information.
The distance from the sensor to the interference object after the deviation can be predicted through the deviation amount and the deviation angle of the trolley:
wherein the content of the first and second substances,
to offset the distance of the sensor to the interfering object,
to offset the distance of the front sensor to the interferer,
for the offset distance to be found from the offset amount,
in order to offset the angle of the angle,
is the angle between the interferent and the sensor.
And further predicting the distance from each sampling point to the interferent after the offset:
wherein the content of the first and second substances,
is the first after offset
The distance of each sample point to the interfering object,
to offset the distance of the sensor to the interfering object,
to offset the distance of the front sensor to the interferer,
in order to offset the angle of the angle,
is the angle between the interferent and the sensor.
Because the farther the sensor is from the interfering object, the less the interference, and the closer the sensor is, the stronger the interference. Therefore, the weight of the sampling point of the sensor on the distance can be obtained
Comprises the following steps:
wherein the content of the first and second substances,
is as follows
The weight of each sample point over the distance,
in order to be the number of sample points,
is the first after offset
Distance of each sample point to an interfering object.
The final weight can be obtained according to the signal value ratio and the weight on the distance:
wherein the content of the first and second substances,
in order to be the final weight, the weight,
is a first
Each sample point corresponds to the total value of the signal values of the cluster,
is a first
The signal values of the individual sampling points are,
as to the number of sampling points,
is as follows
The weight of each sample point over distance.
The weight of each sampling point is distributed, and then the sensitivity of the sensor is adjusted, so that the influence of the trolley on the interferent in the advancing process can work stably.
In summary, in the embodiment of the invention, the RGBD camera deployed on the AGV cart is used to acquire the image of the front view path of the cart, so as to obtain the RGB image and the depth image. And analyzing the interference object on the RGB image, judging whether the interference object on the current path is a metal interference object, and if the interference object is the metal interference object, acquiring interference object information according to the RGB image and the depth image information. And processing the interference object information through an offset prediction network to obtain an offset angle, and correcting the position of the trolley according to the offset angle. Predicting the predicted distance information of each sampling point of the sensor and the interferent after the trolley is corrected, and performing weighting adjustment on the sensitivity of the sampling point of the sensor according to the distance information and the magnetic induction intensity information received by the sampling point, so that the trolley can work stably.
Referring to fig. 2, a block diagram of an AGV car magnetic tape jammer detection system based on artificial intelligence according to an embodiment of the present invention is shown, where the system includes an image acquisition module 101, an image analysis module 102, an offset angle acquisition module 103, and a sensor adjustment module 104.
The image acquisition module 101 is configured to acquire an RGB image and a depth image in a front view path through an RGBD camera mounted on the cart when the magnetic induction signal acquired by the sensor is abnormal.
The image analysis module 102 is configured to divide the RGB image into magnetic tape regions as regions of interest, and perform interferent analysis on the RGB image and the depth image in the regions of interest to obtain interferent images. And obtaining the information of the interferent according to the images of the interferent. The interferent information comprises the area of the interferent, the distance from the interferent to the sensor, the degree of magnetization of the interferent, and the interference mapping range.
The offset angle obtaining module 103 is configured to process the interfering object information by using a pre-trained offset prediction network to obtain an offset angle. And correcting the position of the trolley by utilizing the offset angle.
The sensor adjusting module 104 is configured to predict predicted distance information of each sampling point of the sensor corrected by the cart and the interfering object, and the predicted distance information combines the magnetic induction intensity information received by the sampling point to perform weighting adjustment on the sensitivity of the sampling point.
Preferably, the image analysis module 102 further comprises a semantic segmentation module. The semantic segmentation module is used for extracting the color, luster and texture of the interferent in the interested area as interferent characteristics to judge whether the interferent is a metal interferent. And if the interferent is a non-metal interferent, recording the current position information of the trolley. And if the interference object is a metal interference object, recording the current position information and simultaneously obtaining an interference object image.
Preferably, the image analysis module 102 further comprises an interference mapping range acquisition module. The interference mapping range acquisition module is used for acquiring the length of the interference object according to the coordinate information in the interference object image, and the ratio of the length of the sensor to the length of the interference object is used as an interference mapping range.
Preferably, the sensor adjustment module 104 further comprises a cluster analysis module. The cluster analysis module is used for obtaining a plurality of groups of signal cluster through cluster analysis of the signal value of each sampling point. And calculating the ratio of each cluster of signal values in the total signal value to obtain a signal ratio as magnetic induction intensity information.
Preferably, the system further comprises a tape position correction module. The magnetic tape position correction module is used for determining that the trolley is not positioned on the magnetic tape and correcting the position of the trolley if the magnetic induction intensity signal is smaller than a preset magnetic induction intensity threshold value when the sensor acquires the magnetic induction intensity signal.
It should be noted that: the sequence of the above embodiments of the present invention is only for description, and does not represent the advantages or disadvantages of the embodiments. And specific embodiments thereof have been described above. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims can be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (10)
1. An AGV trolley magnetic tape interferent detection method based on artificial intelligence is characterized by comprising the following steps:
when the magnetic induction intensity signal acquired by the sensor is abnormal, acquiring an RGB image and a depth image under an orthographic path through an RGBD (red, green and blue) camera carried by the trolley;
dividing the RGB image into magnetic band regions as regions of interest, and performing interference analysis on the RGB image and the depth image in the regions of interest to obtain interference images; obtaining interference object information according to the interference object image; the interferent information comprises the area of the interferent, the distance from the interferent to the sensor, the magnetization degree of the interferent and an interference mapping range;
processing the interferent information by utilizing a pre-trained offset prediction network to obtain an offset angle; the offset prediction network is a time domain convolution network; correcting the position of the trolley by using the offset angle;
predicting the predicted distance information of each sampling point of the sensor and the interferent after the trolley is corrected, wherein the predicted distance information combines the magnetic induction intensity information received by the sampling points to perform weighting adjustment on the sensitivity of the sampling points, and the method specifically comprises the following steps:
clustering the sampling signals at each moment to obtain the ratio of each cluster of signals in the total signal value; obtaining the final weight of the magnetic induction intensity of each sampling point through a final weight formula, wherein the final weight formula comprises the following components:
wherein the content of the first and second substances,
in order to be the final weight, the weight,
is as follows
Each sample point corresponds to the total value of the signal values of the cluster,
is a first
The signal values of the individual sampling points are,
in order to be able to count the number of signal points,
is the first after offset
Distance of each sample point to the interferer.
2. The AGV car tape interferent detection method based on artificial intelligence of claim 1, wherein the performing interferent analysis on the RGB images and the depth images within the region of interest comprises:
extracting the color, luster and texture of the interferent in the region of interest as interferent characteristics to judge whether the interferent is a metal interferent; if the interferent is a non-metal interferent, recording the current trolley position information; and if the interference object is a metal interference object, recording the current position information and simultaneously obtaining the image of the interference object.
3. The AGV car magnetic tape interferent detection method based on artificial intelligence of claim 1, wherein the interference mapping range obtaining method comprises:
and obtaining the length of the interference object according to the coordinate information in the interference object image, and taking the ratio of the length of the sensor to the length of the interference object as the interference mapping range.
4. The AGV car magnetic tape interferent detection method based on artificial intelligence of claim 1, wherein the magnetic induction intensity information obtaining method comprises:
obtaining a plurality of groups of signal clusters by clustering and analyzing the signal value of each sampling point; and calculating the ratio of each cluster of signal values in the total signal value to obtain a signal ratio as the magnetic induction intensity information.
5. The AGV car magnetic tape interferent detection method based on artificial intelligence of claim 1, wherein when the sensor collects magnetic induction signals, if the magnetic induction signals are smaller than a preset magnetic induction threshold, it is determined that the car is not on a magnetic tape, and the position of the car is corrected.
6. An AGV dolly magnetic tape interferent detecting system based on artificial intelligence, its characterized in that, this system includes: the device comprises an image acquisition module, an image analysis module, an offset angle acquisition module and a sensor adjustment module;
the image acquisition module is used for acquiring RGB images and depth images under an orthographic path through an RGBD (red, green and blue) camera carried by the trolley when magnetic induction intensity signals acquired by the sensor are abnormal;
the image analysis module is used for dividing the RGB image into magnetic tape regions as regions of interest, and performing interference object analysis on the RGB image and the depth image in the regions of interest to obtain interference object images; obtaining interference object information according to the interference object image; the interferent information comprises the area of the interferent, the distance from the interferent to the sensor, the magnetization degree of the interferent and an interference mapping range;
the offset angle acquisition module is used for processing the interferent information by using a pre-trained offset prediction network to acquire an offset angle; the offset prediction network is a time domain convolution network; correcting the position of the trolley by using the offset angle;
the sensor adjusting module is used for predicting the predicted distance information of each sampling point of the sensor and the interferent after the trolley is corrected, and the predicted distance information combines the magnetic induction intensity information received by the sampling points to perform weighting adjustment on the sensitivity of the sampling points, and specifically comprises the following steps:
clustering the sampling signals at each moment to obtain the ratio of each cluster of signals in the total signal value; obtaining the final weight of the magnetic induction intensity of each sampling point through a final weight formula, wherein the final weight formula comprises the following components:
wherein the content of the first and second substances,
in order to be the final weight, the weight,
is as follows
The individual sample points correspond to the total value of the signal values of the cluster,
is as follows
The signal values of the individual sampling points are,
in order to be able to count the number of signal points,
is the first after offset
Distance of each sample point to an interfering object.
7. The AGV car magnetic tape interferent detection system based on artificial intelligence of claim 6, wherein the image analysis module further comprises a semantic segmentation module;
the semantic segmentation module is used for extracting the color, luster and texture of the interferent in the region of interest as interferent characteristics to judge whether the interferent is a metal interferent; if the interferent is a non-metal interferent, recording the current position information of the trolley; and if the interference object is a metal interference object, recording the current position information and simultaneously obtaining the image of the interference object.
8. The AGV car magnetic tape interferent detection system based on artificial intelligence of claim 6, wherein the image analysis module further comprises an interference mapping range acquisition module;
the interference mapping range acquisition module is used for acquiring the length of the interference object according to the coordinate information in the interference object image, and the ratio of the length of the sensor to the length of the interference object is used as the interference mapping range.
9. The AGV car tape interferers detection system of claim 6, wherein the sensor adjustment module further comprises a cluster analysis module;
the cluster analysis module is used for obtaining a plurality of groups of signal clusters by clustering and analyzing the signal value of each sampling point; and calculating the ratio of each cluster of signal values in the total signal value to obtain a signal ratio as the magnetic induction intensity information.
10. The AGV car tape interferent detection system based on artificial intelligence of claim 6, further comprising a tape position correction module;
the magnetic tape position correction module is used for determining that the trolley is not positioned on the magnetic tape and correcting the position of the trolley if the magnetic induction intensity signal is smaller than a preset magnetic induction intensity threshold value when the sensor acquires the magnetic induction intensity signal.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110412844.7A CN113076910B (en) | 2021-04-16 | 2021-04-16 | AGV trolley magnetic tape interferent detection method and system based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110412844.7A CN113076910B (en) | 2021-04-16 | 2021-04-16 | AGV trolley magnetic tape interferent detection method and system based on artificial intelligence |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113076910A CN113076910A (en) | 2021-07-06 |
| CN113076910B true CN113076910B (en) | 2022-12-13 |
Family
ID=76617801
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110412844.7A Active CN113076910B (en) | 2021-04-16 | 2021-04-16 | AGV trolley magnetic tape interferent detection method and system based on artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113076910B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101387522A (en) * | 2008-09-02 | 2009-03-18 | 吉林大学 | Magnetic guide sensor |
| CN106370189A (en) * | 2016-12-02 | 2017-02-01 | 华中科技大学 | Multi-sensor fusion-based indoor navigation device and method |
| CN207029364U (en) * | 2017-08-14 | 2018-02-23 | 哈尔滨博乐恩机器人技术有限公司 | AGV dollies with anti-magnetic disturbance |
| CN108762276A (en) * | 2018-06-07 | 2018-11-06 | 安徽理工大学 | A kind of automatic inclined rail means for correcting of AGV trolleies and automatic rail bearing calibration partially |
| CN110209170A (en) * | 2019-06-21 | 2019-09-06 | 珠海丽亭智能科技有限公司 | A kind of travel track antidote for the robot that stops |
| CN110320906A (en) * | 2019-05-09 | 2019-10-11 | 重庆大学 | A kind of 4 wheel driven AGV trolley differential straight-line travelling attitude adjusting method based on Mecanum wheel |
| CN111474933A (en) * | 2020-04-24 | 2020-07-31 | 合肥工业大学 | Automatic deviation rectification control method of magnetic guidance AGV |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2007219960A (en) * | 2006-02-20 | 2007-08-30 | Tcm Corp | Position deviation detection device |
| BR102013010820A2 (en) * | 2013-05-02 | 2015-06-30 | Kelbe Participaç Es Ltda | Sugar cane cutter |
| CN211718764U (en) * | 2020-04-14 | 2020-10-20 | 辽宁正集电气技术有限公司 | Magnetic navigation sensing device capable of dynamically adjusting sensitivity |
| CN112330695A (en) * | 2020-11-26 | 2021-02-05 | 河南耀蓝智能科技有限公司 | Automatic water adding and changing decision method and system based on artificial intelligence |
-
2021
- 2021-04-16 CN CN202110412844.7A patent/CN113076910B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101387522A (en) * | 2008-09-02 | 2009-03-18 | 吉林大学 | Magnetic guide sensor |
| CN106370189A (en) * | 2016-12-02 | 2017-02-01 | 华中科技大学 | Multi-sensor fusion-based indoor navigation device and method |
| CN207029364U (en) * | 2017-08-14 | 2018-02-23 | 哈尔滨博乐恩机器人技术有限公司 | AGV dollies with anti-magnetic disturbance |
| CN108762276A (en) * | 2018-06-07 | 2018-11-06 | 安徽理工大学 | A kind of automatic inclined rail means for correcting of AGV trolleies and automatic rail bearing calibration partially |
| CN110320906A (en) * | 2019-05-09 | 2019-10-11 | 重庆大学 | A kind of 4 wheel driven AGV trolley differential straight-line travelling attitude adjusting method based on Mecanum wheel |
| CN110209170A (en) * | 2019-06-21 | 2019-09-06 | 珠海丽亭智能科技有限公司 | A kind of travel track antidote for the robot that stops |
| CN111474933A (en) * | 2020-04-24 | 2020-07-31 | 合肥工业大学 | Automatic deviation rectification control method of magnetic guidance AGV |
Non-Patent Citations (5)
| Title |
|---|
| Design of magnetic navigation automatic guided vehicle system;Ruwen Chen et al;《2019 2nd International Conference on Clean Energy and Electrical Systems》;20191231;第1-8页 * |
| Neural Network Control for Automatic Guided Vehicles Using Discrete Reference Markers;Saed Kurd et al;《IAS 97. Conference Record of the 1997 IEEE Industry Applications Conference Thirty-Second IAS Annual Meeting》;20020806;第886-891页 * |
| 基于双目视觉的AGV障碍物检测与避障;王铮等;《计算机集成制造系统》;20180215(第02期);第400-409页 * |
| 基于彩色轨迹引导的AGV视觉导航方案研究;龙水军等;《工具技术》;20140120(第01期);第74-78页 * |
| 基于最优偏差路径的自动导引车纠偏方法;罗哉等;《仪器仪表学报》;20170415(第04期);第853-860页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113076910A (en) | 2021-07-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Sampedro et al. | A supervised approach to electric tower detection and classification for power line inspection | |
| CN104200657B (en) | A kind of traffic flow parameter acquisition method based on video and sensor | |
| CN104282020A (en) | Vehicle speed detection method based on target motion track | |
| CN106897681B (en) | Remote sensing image contrast analysis method and system | |
| CN112364768B (en) | Distributed optical fiber intrusion recognition method based on airspace characteristics and machine learning | |
| CN107665348B (en) | Digital identification method and device for digital instrument of transformer substation | |
| CN104361340A (en) | SAR image target fast detecting method based on significance detecting and clustering | |
| CN107516423B (en) | Video-based vehicle driving direction detection method | |
| CN112927303B (en) | Lane line-based automatic driving vehicle-mounted camera pose estimation method and system | |
| CN102214290B (en) | License plate positioning method and license plate positioning template training method | |
| CN112800938B (en) | Method and device for detecting occurrence of side rockfall of unmanned vehicle | |
| CN105444741A (en) | Double view window based route characteristic identifying, deviation measuring, and accurate positioning method | |
| CN105139391A (en) | Edge detecting method for traffic image in fog-and-haze weather | |
| Zhou et al. | Autonomous detection of crop rows based on adaptive multi-ROI in maize fields. | |
| CN113076910B (en) | AGV trolley magnetic tape interferent detection method and system based on artificial intelligence | |
| CN111950498A (en) | Lane line detection method and device based on end-to-end instance segmentation | |
| CN111367901A (en) | Ship data denoising method | |
| CN112129290A (en) | System and method for monitoring riding equipment | |
| CN114609609A (en) | Speed estimation method for extracting static point cloud by FMCW laser radar random sampling | |
| CN111717210B (en) | Detection method for separation of driver from steering wheel in relative static state of hands | |
| CN113221739A (en) | Monocular vision-based vehicle distance measuring method | |
| JP5946294B2 (en) | Object detection device, object detection method, object detection program, and autonomous vehicle | |
| CN116309407A (en) | Method for detecting abnormal state of railway contact net bolt | |
| CN116188943A (en) | Solar radio spectrum burst information detection method and device | |
| CN115880674A (en) | Obstacle avoidance and steering correction method based on unmanned mine car |
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 | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20221128 Address after: 710000 Room 206, Building 2, Hangchuang International Plaza, Shenzhou 4th Road, Xi'an National Civil Aerospace Industry Base, Shaanxi Province Applicant after: Shaanxi hengtuochuangcheng Technology Co.,Ltd. Address before: 471000 Room 502, building 17, area C, Shenglong square, Xigong District, Luoyang City, Henan Province Applicant before: Henan Liumi Electronic Technology Co.,Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |