CN115410150A

CN115410150A - Detection method and detection device for deviation of conveyor belt and processor

Info

Publication number: CN115410150A
Application number: CN202211008458.2A
Authority: CN
Inventors: 武国平; 张琦; 叶军; 李志军; 李�昊; 吉日格勒
Original assignee: China Shenhua Energy Co Ltd
Current assignee: China Shenhua Energy Co Ltd
Priority date: 2022-08-22
Filing date: 2022-08-22
Publication date: 2022-11-29

Abstract

The application provides a detection method, a detection device and a processor for deviation of a conveyor belt. The method comprises the following steps: acquiring historical image data of a conveyor belt; constructing a neural network model; acquiring current image data of a conveyor belt; and detecting the current image data by adopting a neural network model, and determining whether the conveying belt deviates. In the scheme, a new neural network model is constructed, the calculation efficiency of the model is higher, the edge of the conveying belt is detected by adopting the new neural network model, the novel neural network model can adapt to a complex environment, the calculation capability is higher than that of a traditional algorithm, the calculation accuracy is higher, the connectivity of the edge of the conveying belt is better, the edge of the conveying belt can be stably detected in an actual environment, whether the conveying belt deviates or not is accurately determined through the detected edge, and therefore the efficiency of detecting whether the conveying belt deviates or not is improved.

Description

Detection method and detection device for deviation of conveyor belt and processor

Technical Field

The application relates to the field of image processing, in particular to a detection method and a detection device for conveyor belt deviation, a computer readable storage medium and a processor.

Background

The deviation of the conveying belt is one of the highest faults of the belt conveyor, which can cause the materials conveyed on the conveying belt to be scattered and also can influence the safe operation of conveying equipment. The method is characterized in that the working video of a conveyor belt in a mine tunnel of a coal preparation plant is collected in real time at present and is monitored manually and remotely, or the conveyor belt is manually observed on site, the manual detection method is time-consuming and labor-consuming and has low detection efficiency, and the method is also characterized in that whether the conveyor belt deviates or not is detected by adopting a method of combining image recognition technology and edge detection algorithm detection, but the detection efficiency is low due to the fact that the detection speed of the current detection algorithm is low and the detection time is long.

Disclosure of Invention

The application mainly aims to provide a detection method, a detection device, a computer readable storage medium and a processor for detecting whether a conveyor belt deviates, so as to solve the problem that the efficiency of detecting whether the conveyor belt deviates in the prior art is low.

According to an aspect of the embodiments of the present invention, there is provided a method for detecting deviation of a conveyor belt, including: acquiring historical image data of a conveyor belt, wherein the historical image data is acquired by image acquisition equipment which is arranged above the conveyor belt; constructing a neural network model, wherein the neural network is obtained by training a plurality of groups of training data, and each group of training data in the plurality of groups of training data comprises the historical image data and the related information of the conveyor belt corresponding to the historical image data; acquiring current image data of the conveyor belt; and detecting the current image data by adopting the neural network model to determine whether the conveyor belt deviates.

Optionally, constructing the neural network model comprises: acquiring target pixel points and non-target pixel points in a target area, wherein the target pixel points refer to central pixel points in the target area, the non-target pixel points refer to pixel points in the target area except the target pixel points, and the target area refers to an area including the edge of the conveyor belt in a historical image; determining a first area and a second area, wherein the first area is an area in which the brightness difference between the target pixel point and the non-target pixel point is greater than or equal to a brightness threshold, and the second area is an area in which the brightness difference between the target pixel point and the non-target pixel point is less than the brightness threshold; respectively carrying out operation on the image data of the first area and the image data of the second area and a kernel function to obtain first brightness data and second brightness data; and inputting the first brightness data and the second brightness data into a radial basis function neural network model of Gaussian projection for operation, wherein the radial basis function neural network model of Gaussian projection comprises an input layer, an expansion layer, a hidden layer and an output layer, the input layer is used for receiving input data, the expansion layer is used for performing difference on two adjacent input data, the hidden layer is used for calculating the input data subjected to difference and a Gaussian function, and the output layer is used for outputting a calculation result, wherein the input data is an absolute value of a product of the first brightness data and the second brightness data.

Optionally, after constructing the neural network model, the method further comprises: determining a central point, wherein the central point refers to the input data of the hidden layer of the neural network model; acquiring the maximum value and the minimum value of the pixel values of the input data and the number of the pixel points of the input data; determining a variance by adopting an implicit function according to the maximum value of the pixel values, the minimum value of the pixel values and the number of the pixel points; and calculating an output weight of the neural network model by adopting a least square method, wherein the output weight refers to a weight value occupied by the output data of the hidden layer in the calculation result of the output layer.

Optionally, after constructing the neural network model, the method further comprises: acquiring standard image data; calculating an error value between a calculation result output by the neural network model and the standard image data; determining that the neural network model has been trained to end if the error value is less than an error threshold; and under the condition that the error value is greater than or equal to the error threshold value, determining that the neural network model needs to be retrained, and retraining the neural network model.

Optionally, the detecting the current image data by using the neural network model to determine whether the conveyor belt is off tracking includes: determining a first edge position and a second edge position of the current image data by using the neural network model; acquiring a standard center line position of the conveyor belt, wherein the standard center line position refers to a center line position when the conveyor belt is not deviated; and determining whether the conveyor belt deviates or not according to the first edge position, the second edge position and the standard center line position.

Optionally, determining whether the conveyor belt is off tracking according to the first edge position, the second edge position and the standard center line position includes: calculating a first position difference between the first edge position and the standard centerline position in real time; calculating a second position difference value of the second edge position and the standard centerline position in real time; determining that the conveyor belt has deviated when the first position difference is less than or equal to a position threshold, or the second position difference is less than or equal to the position threshold; and determining that the conveyor belt is not off tracking under the condition that the first position difference value is greater than the position threshold value and the second position difference value is greater than the position threshold value.

Optionally, in the case that it is determined that the conveyor belt has deviated, the method further comprises: generating alarm information, wherein the alarm information is used for prompting that the conveyor belt is off tracking; and controlling the conveyor belt to pause.

According to another aspect of the embodiments of the present invention, there is also provided a detection apparatus for deviation of a conveyor belt, including: the device comprises a first acquisition unit, a second acquisition unit and a control unit, wherein the first acquisition unit is used for acquiring historical image data of a conveyor belt, the historical image data is acquired by an image acquisition device, and the image acquisition device is arranged above the conveyor belt; the building unit is used for building a neural network model, wherein the neural network is obtained by using a plurality of groups of training data through training, and each group of training data in the plurality of groups of training data comprises the historical image data and the related information of the conveyor belt corresponding to the historical image data; a second acquisition unit configured to acquire current image data of the conveyor belt; and the detection unit is used for detecting the current image data by adopting the neural network model and determining whether the conveying belt deviates.

According to still another aspect of embodiments of the present invention, there is also provided a computer-readable storage medium including a stored program, wherein the program executes any one of the methods.

According to still another aspect of the embodiments of the present invention, there is further provided a processor, configured to execute the program, where the program executes to perform any one of the methods.

In the embodiment of the invention, the historical image data of the conveyor belt is firstly acquired, then the neural network model is constructed, then the current image data of the conveyor belt is acquired, and finally the neural network model is adopted to detect the current image data to determine whether the conveyor belt deviates. In the scheme, a new neural network model is constructed, the calculation efficiency of the model is higher, the new neural network model is adopted to detect the edges of the conveying belt, the novel neural network model can adapt to a complex environment, the calculation capability is higher than that of a traditional algorithm, the calculation accuracy is higher, the connectivity of the edges of the conveying belt is better, the edges of the conveying belt can be stably detected in an actual environment, whether the conveying belt deviates or not is accurately determined through the detected edges, and therefore the efficiency of detecting whether the conveying belt deviates or not is improved.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this application, illustrate embodiments of the application and, together with the description, serve to explain the application and are not intended to limit the application. In the drawings:

fig. 1 shows a schematic flow chart of a method for detecting deviation of a conveyor belt according to an embodiment of the application;

FIG. 2 shows a schematic structural diagram of a constructed neural network model;

FIG. 3 shows a schematic of the results of the algorithmic processing;

FIG. 4 shows a schematic structural diagram of a conveyor belt deviation detection device according to an embodiment of the application;

fig. 5 shows a schematic flow chart of another detection method for deviation of a conveyor belt according to an embodiment of the application.

Wherein the figures include the following reference numerals:

100. a conveyor belt; 200. a first edge position; 300. a second edge position; 400. standard centerline position.

Detailed Description

It should be noted that the embodiments and features of the embodiments in the present application may be combined with each other without conflict. The present application will be described in detail below with reference to the embodiments with reference to the attached drawings.

In order to make the technical solutions better understood by those skilled in the art, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only partial embodiments of the present application, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.

It should be noted that the terms "first," "second," and the like in the description and claims of this application and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It should be understood that the data so used may be interchanged under appropriate circumstances such that embodiments of the application described herein may be used. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.

It will be understood that when an element such as a layer, film, region, or substrate is referred to as being "on" another element, it can be directly on the other element or intervening elements may also be present. Also, in the specification and claims, when an element is described as being "connected" to another element, the element may be "directly connected" to the other element or "connected" to the other element through a third element.

Image edge detection is a main problem in image processing and computer vision, and a large number of researches show that the edge detection plays an important role in the aspects of image high-order feature extraction, feature description, target identification, image segmentation and the like, how to quickly and accurately position and extract image edge feature information becomes one of research hotspots, edge detection algorithms are rich, and the method can be roughly divided into edge detection algorithms based on gradients, edge detection algorithms based on mathematical morphology, edge detection algorithms based on genetic algorithms and edge detection algorithms based on neural networks by fusing new theories.

For processing the working video of the conveyor belt in the mine tunnel of the coal preparation plant in real time, on one hand, the algorithm is required to be capable of adapting to a relatively complex environment, on the other hand, the algorithm is required to consume low computing resources, and real-time processing can be realized when the real-time monitoring video is received.

Because the image Edge contains a large amount of background information and important structural information, the traditional Edge detection method usually uses the bottom layer characteristics of manual manufacture such as color, brightness, gradient and the like as the priority of Edge detection, such as a Sobel operator, a coefficient Code gradient (SCG for short) and a Structured forest Edge detection (SE for short) algorithm. Although the edge detection method using low-level features has been greatly developed, its limitations are also obvious, and with the development of deep learning technology, especially the appearance of Convolutional Neural Network (CNN), people find that the Neural Network has strong automatic learning capability in the natural image field, but the edge detection method based on CNN relies on a classification Network, and because the classification method is slow, the detection time is long, the detection is not flexible, the noise is also large, and the detection efficiency is not high.

As mentioned in the background of the invention, in order to solve the above problem, the prior art has low efficiency of detecting whether the conveyor belt is off tracking, and in an exemplary embodiment of the present application, a method, an apparatus, a computer-readable storage medium, and a processor for detecting off tracking of the conveyor belt are provided.

According to the embodiment of the application, a method for detecting deviation of a conveyor belt is provided.

Fig. 1 is a flowchart of a method for detecting deviation of a conveyor belt according to an embodiment of the present application. As shown in fig. 1, the method comprises the steps of:

step S101, acquiring historical image data of a conveyor belt, wherein the historical image data is acquired by image acquisition equipment which is arranged above the conveyor belt;

in particular, the image acquisition device may be a camera, video camera, still camera, scanner, or the like.

Step S102, a neural network model is constructed, wherein the neural network is obtained by using a plurality of groups of training data for training, and each group of training data in the plurality of groups of training data comprises the historical image data and the related information of the conveyor belt corresponding to the historical image data;

step S103, acquiring current image data of the conveyor belt;

and step S104, detecting the current image data by adopting the neural network model, and determining whether the conveying belt deviates.

In the method, the historical image data of the conveyor belt is firstly obtained, then the neural network model is constructed, then the current image data of the conveyor belt is obtained, and finally the neural network model is adopted to detect the current image data to determine whether the conveyor belt deviates. In the scheme, a new neural network model is constructed, the calculation efficiency of the model is higher, the edge of the conveying belt is detected by adopting the new neural network model, the novel neural network model can adapt to a complex environment, the calculation capability is higher than that of a traditional algorithm, the calculation accuracy is higher, the connectivity of the edge of the conveying belt is better, the edge of the conveying belt can be stably detected in an actual environment, whether the conveying belt deviates or not is accurately determined through the detected edge, and therefore the efficiency of detecting whether the conveying belt deviates or not is improved.

It should be noted that the steps illustrated in the flowcharts of the figures may be performed in a computer system such as a set of computer-executable instructions and that, although a logical order is illustrated in the flowcharts, in some cases, the steps illustrated or described may be performed in an order different than here.

In one embodiment, the neural network model may be a gaussian projected radial basis neural network model, and in particular, the image acquisition device may sample the belt-fed image data at predetermined time intervals. The collected historical images are preprocessed, and because the environment of the coal preparation plant is dark relatively, and the noise points of image data are more, gaussian filtering is adopted to smooth the images. The Gaussian filter calculates the weighted average value of the original image, namely the value of each pixel point, and can be obtained by calculating the original value of the pixel point and the value of the adjacent pixel point through the weighted average, each pixel point is sequentially traversed during calculation, the pixel point calculated in each round is taken as a central point (0, 0), the sigma multiplied by sigma pixel points in adjacent squares are considered, and the weighted value of each pixel point is determined by a Gaussian function:

wherein x represents the abscissa of the pixel point relative to the central point, y represents the ordinate of the pixel point relative to the central point, σ represents the width of the gaussian function, G (x, y) represents the gaussian function of the pixel point, i.e. the weight of the position in the gaussian filtering process, σ × σ pixels in the width range are weighted and averaged according to the weight value of each pixel point, and the obtained final value is the pixel gray value of the pixel point after gaussian filtering. And when all the pixels are traversed, the Gaussian filtering of the image can be completed through calculation, and the preprocessing is completed. And then inputting each pixel point in the preprocessed image into a radial basis function neural network based on Gaussian projection.

In one embodiment of the present application, constructing a neural network model includes: acquiring target pixel points and non-target pixel points in a target area, wherein the target pixel points refer to central pixel points in the target area, the non-target pixel points refer to pixel points in the target area except the target pixel points, and the target area refers to an area including the edge of the conveyor belt in a historical image; determining a first region and a second region, wherein the first region is a region in which a brightness difference between the target pixel point and the non-target pixel point is greater than or equal to a brightness threshold, and the second region is a region in which the brightness difference between the target pixel point and the non-target pixel point is less than the brightness threshold; respectively calculating the image data of the first area and the image data of the second area with a kernel function to obtain first brightness data and second brightness data; and inputting the first luminance data and the second luminance data into a radial basis function neural network model of gaussian projection for operation, wherein the radial basis function neural network model of gaussian projection comprises an input layer, an expansion layer, a hiding layer and an output layer, the input layer is used for receiving input data, the expansion layer is used for performing difference on two adjacent input data, the hiding layer is used for calculating the input data after difference and a gaussian function, and the output layer is used for outputting a calculation result, wherein the input data is an absolute value of a product of the first luminance data and the second luminance data. In the embodiment, the calculation efficiency of the radial basis function neural network model of the Gaussian projection is higher, the calculated edge connectivity is better, the information of the edge of the conveyor belt can be analyzed more efficiently and accurately, and the detection efficiency can be further improved.

For human vision, when the brightness jumps, there is a perception of edge enhancement that the bright side is visually perceived as brighter and the dark side is visually perceived as darker, a characteristic known as the mach effect. With the mach effect, the following two mathematical expressions can be proposed:

(1) And each pixel point is subjected to weighted comparison of distance with surrounding points, namely:

(2) When a certain pixel point is compared, the pixel value of the surrounding points is judged to be higher than that of the central point, namely:

wherein, D represents the whole image domain (target region), D1 represents the first region where the difference between the target pixel point x and the non-target pixel point y is larger than zero, D2 represents the second region where the difference between the pixel point x and the non-target pixel point y is smaller than zero, I (-) represents the pixel value of the point, K (-) is a Gaussian kernel function, mid (-) represents the difference between the target pixel point x and the non-target pixel point y, and GP (Gaussian Positive) and GN (Gaussian Negative) are respectively the Positive and Negative parts of the difference between the target pixel point x and the non-target pixel point y, namely the first brightness data and the second brightness data.

From the above, the GP and GN values can be obtained, and the edge contour highlighting can be further realized by the following steps:

(1) Letting n =1, taking a pixel xn in the image with h pixels and a 3 × 3 region thereof, and calculating the difference between the pixel values of the xn and the non-target pixels to obtain regions D1 and D2;

(2) Multiplying the values of the two regions by a Kernel function Kernel respectively and adding the values to obtain GP values and GN values respectively;

(3) The value of xn is reassigned to-GN × GP;

(4) And if n is less than h, adding 1 to n, taking the next pixel point, and repeating until n = h.

The above process can be understood as a gaussian projection process, a gaussian projection radial basis function neural network model is further constructed to achieve the purpose of highlighting the contour, the structure diagram of the model is shown in fig. 2, an input layer consists of pixel points xn in an image with h pixel points, the difference value of the pixel values of the xn and non-target pixel points is calculated, the result is input to a second layer expansion layer, a third layer is a hidden layer, the hidden layer uses a gaussian function as a hidden function, and the hidden function is:

wherein x is the output result of the previous layer, the parameter represents the variance of the gaussian function, the weight values of all nodes in the hidden layer are the same, and the weights of all nodes except the hidden layer in the radial basis function neural network model of the gaussian projection are all 1.

In another embodiment of the present application, after the constructing the neural network model, the method further includes: determining a central point, wherein the central point refers to the input data of the hidden layer of the neural network model; acquiring the maximum value and the minimum value of the pixel values of the input data and the number of the pixel points of the input data; determining a variance by adopting an implicit function according to the maximum value of the pixel values, the minimum value of the pixel values and the number of the pixel points; and calculating an output weight of the neural network model by adopting a least square method, wherein the output weight refers to a weight value occupied by the output data of the hidden layer in the calculation result of the output layer. In the scheme, after the neural network model is constructed, the neural network is further trained, so that the trained neural network model is higher in calculation efficiency.

Specifically, the node of the hidden layer is d _i I =1,2, \ 8230h, h is the total number of input image data, and the corresponding center point is made to be each input data, so that the input data can be directly mapped to the hidden layer.

Specifically, the number of nodes of the hidden layer is the same as the number of input data, and in order to avoid loss of image information, the central point is set as each input data according to the above formula, so that the integrity of the image information is ensured, and the efficiency of the model can be further improved.

The neural network model uses a Gaussian function as an implicit function, and the variance is calculated by the formula:

sigma represents variance, xmax represents the maximum value of a pixel value, xmin represents the minimum value of the pixel value, and h represents the number of pixel points, so that a mode of directly solving the variance can obtain more appropriate variance for different images, and the training efficiency is accelerated to a certain extent.

Since the number of nodes and the variance in the neural network model are determined, the input to the output can be represented by a linear equation, and the output weight can be solved by a least square method, wherein the formula is as follows:

i =1,2, a., h, P =1,2, a., P, where ω denotes an output weight,

representing the P-th input image, P =1,2, \ 8230;, P, P being the total number of input images, d _i Representing the respective center point, i.e. the respective node of the hidden layer.

The radial basis function neural network of Gaussian projection adopts the characteristics of sparse connection and weight sharing in CNN, reduces the complexity of a neural network model, can directly apply the neural network model to an image with multidimensional input vectors, and reduces the complexity of feature extraction. Therefore, the neural network model can be regarded as an extension of the radial basis function neural network, a fixed weight is added to the first layer, and then the filtered image forms an input sample which is used as an input layer to train the Gaussian projection radial basis function neural network.

In another embodiment of the present application, after the building the neural network model, the method further includes: acquiring standard image data; calculating an error value between a calculation result output by the neural network model and the standard image data; determining that the neural network model is trained and ended when the error value is smaller than an error threshold value; and determining that the neural network model needs to be retrained when the error value is greater than or equal to the error threshold value, and retraining the neural network model. In this embodiment, the calculated error value of the neural network model is detected, and the neural network model is retrained under the condition of a large error value, so that the efficiency of the neural network model can be further ensured to be good.

In another embodiment of the present application, detecting the current image data by using the neural network model to determine whether the conveyor belt is off-tracking includes: determining a first edge position and a second edge position of the current image data by adopting the neural network model; acquiring a standard center line position of the conveyor belt, wherein the standard center line position refers to a center line position when the conveyor belt is not deviated; and determining whether the conveyor belt deviates according to the first edge position, the second edge position and the standard center line position. In this embodiment, whether the conveyor belt has deviated or not can be further accurately determined by counting the first edge position and the second edge position of the processed image and then according to the first edge position, the second edge position, and the standard center line position.

In a specific embodiment of the present application, determining whether the conveyor belt is off-tracking according to the first edge position, the second edge position and the standard center line position includes: calculating a first position difference between the first edge position and the standard centerline position in real time; calculating a second position difference between the second edge position and the standard centerline position in real time; determining that the conveyor belt has deviated when the first position difference is less than or equal to a position threshold, or the second position difference is less than or equal to the position threshold; and determining that the conveyor belt is not off tracking when the first position difference is greater than the position threshold and the second position difference is greater than the position threshold. In this embodiment, when the difference between the distance between the first edge position or the distance between the second edge position and the standard center line position of the conveyor belt is greater than the position threshold, it indicates that the conveyor belt may not be deflected at this time, or the degree of deflection is not great, and there is no great influence on the conveyor belt, and if the difference between the distance between the first edge position or the distance between the second edge position and the standard center line position is less than or equal to the position threshold, it indicates that the conveyor belt is deflected at this time, and the edge of the conveyor belt may approach the standard center line position, so that this embodiment can further accurately determine whether the conveyor belt is deflected.

In practical application, the edge contour can be accurately determined by the collected current image data and the neural network model, the edge can be extracted by using a contour tracking algorithm according to the output result so as to achieve the purpose of edge detection, if the edge positions in the continuous 5-frame images exceed the position of the median line, the driving belt is proved to be off-tracking, and early warning or alarm processing is needed subsequently.

Specifically, as shown in fig. 3, the result of the algorithm processing is schematically shown, the camera sensor is installed above the conveyor belt 100 of the mine of the coal preparation plant, the camera direction is along the working direction of the conveyor belt 100, the result of the algorithm processing can determine the first edge position 200 and the second edge position 300 of the conveyor belt 100, it can be determined that the conveyor belt 100 is working normally at present, neither the first edge position 200 nor the second edge position 300 reaches the standard center line position 400, and if the first edge position 200 and the second edge position 300 reach the standard center line position 400, a deviation warning is sent.

In another specific embodiment of the present application, in the case that it is determined that the conveyor belt has deviated, the method further includes: generating alarm information, wherein the alarm information is used for prompting that the conveyor belt deviates; and controlling the conveyor belt to pause. In this embodiment, through production alarm information, can in time indicate staff's conveyer belt off tracking this moment, control conveyer belt pause work again, can avoid the conveyer belt off tracking and the dangerous accident that takes place.

Image edge detection plays a crucial role in the fields of image processing, computer vision, and the like. When the image of the mine conveyor belt is processed, the algorithm in the scheme combines the characteristics of a human visual system, and has the capability of adapting to a relatively complex environment, which is a remarkable advantage compared with the traditional edge detection algorithm. Compared with the classic CNN convolutional neural network edge detection method, the method has the advantages that the calculation resources consumed by the algorithm are low, the detection time is short, and meanwhile, a better edge detection result is obtained when an image containing more Gaussian noise is processed, so that the method has a good advantage in a relatively dark coal preparation plant environment with more image noise points, and the conveying belt edge can be stably detected.

And based on the obtained image edge data of each frame, whether the working position of the conveyor belt is normal or not can be quickly judged by detecting the distance between the edge pixel point of the conveyor belt and the standard central line position of the image. Meanwhile, deviation warning improves the stability of system operation through multi-frame error accumulation.

The embodiment of the application also provides a detection device for the deviation of the conveyor belt, and it needs to be explained that the detection device for the deviation of the conveyor belt in the embodiment of the application can be used for executing the detection method for the deviation of the conveyor belt in the embodiment of the application. The following introduces a detection device for deviation of a conveyor belt provided by the embodiment of the application.

Fig. 4 is a schematic diagram of a detection device for deviation of a conveyor belt according to an embodiment of the application. As shown in fig. 4, the apparatus includes:

a first acquisition unit 10, configured to acquire historical image data of a conveyor belt, where the historical image data is acquired by an image acquisition device, and the image acquisition device is installed above the conveyor belt;

A constructing unit 20, configured to construct a neural network model, where the neural network is obtained by using multiple sets of training data, and each set of training data in the multiple sets of training data includes the historical image data and the related information of the conveyor belt corresponding to the historical image data;

a second acquiring unit 30 for acquiring current image data of the conveyor belt;

and the detection unit 40 is used for detecting the current image data by adopting the neural network model and determining whether the conveying belt deviates.

In the scheme, the first acquisition unit acquires historical image data of the conveyor belt, the construction unit constructs a neural network model, the second acquisition unit acquires current image data of the conveyor belt, and the detection unit detects the current image data by adopting the neural network model to determine whether the conveyor belt deviates. In the scheme, a new neural network model is constructed, the calculation efficiency of the model is higher, the edge of the conveying belt is detected by adopting the new neural network model, the novel neural network model can adapt to a complex environment, the calculation capability is higher than that of a traditional algorithm, the calculation accuracy is higher, the connectivity of the edge of the conveying belt is better, the edge of the conveying belt can be stably detected in an actual environment, whether the conveying belt deviates or not is accurately determined through the detected edge, and therefore the efficiency of detecting whether the conveying belt deviates or not is improved.

In one embodiment, the neural network model may be a radial basis neural network model of gaussian projection, and in particular, the image acquisition device may sample the image data of the conveyor belt operation at predetermined time intervals. The collected historical images are preprocessed, and because the environment of the coal preparation plant is relatively dark and the noise of image data is more, the images can be smoothed by Gaussian filtering. The Gaussian filter calculates the weighted average value of the original image, namely the value of each pixel point, can be obtained by calculating the original value of the pixel point and the value of the adjacent pixel point through the weighted average, each pixel point is sequentially traversed during calculation, the pixel point calculated in each round is taken as a central point (0, 0), the sigma multiplied by sigma pixel points in the adjacent checks are considered, and the weighted value of each pixel point is determined by a Gaussian function:

wherein x represents the abscissa of the pixel point relative to the central point, y represents the ordinate of the pixel point relative to the central point, σ represents the width of the gaussian function, G (x, y) represents the gaussian function of the pixel point, i.e. the weight of the position in the gaussian filtering process, σ × σ pixels within the width range are weighted and averaged according to the weight value of each pixel point, and the obtained final value is the pixel gray value of the pixel point after gaussian filtering. And when all the pixels are traversed, the Gaussian filtering of the image can be completed through calculation, and the preprocessing is completed. And then inputting each pixel point in the preprocessed image into a radial basis function neural network based on Gaussian projection.

In an embodiment of the application, the construction unit includes a first obtaining module, a first determining module, an operation module and a processing module, the first obtaining module is configured to obtain a target pixel point and a non-target pixel point in a target region, the target pixel point refers to a central pixel point in the target region, the non-target pixel point refers to a pixel point in the target region other than the target pixel point, and the target region refers to a region including an edge of the conveyor belt in a history image; the first determining module is configured to determine a first region and a second region, where the first region is a region where a luminance difference between the target pixel and the non-target pixel is greater than or equal to a luminance threshold, and the second region is a region where the luminance difference between the target pixel and the non-target pixel is less than the luminance threshold; the operation module is used for respectively operating the image data of the first area, the image data of the second area and a kernel function to obtain first brightness data and second brightness data; the processing module is configured to input the first luminance data and the second luminance data into a radial basis function neural network model of gaussian projection for operation, where the radial basis function neural network model of gaussian projection includes an input layer, an expansion layer, a hidden layer, and an output layer, the input layer is configured to receive input data, the expansion layer is configured to perform a difference between two adjacent input data, the hidden layer is configured to calculate a gaussian function with respect to the input data after the difference is performed, and the output layer is configured to output a calculation result, where the input data is an absolute value of a product of the first luminance data and the second luminance data. In the embodiment, the calculation efficiency of the radial basis function neural network model of the Gaussian projection is higher, the calculated edge connectivity is better, the information of the edge of the conveyor belt can be analyzed more efficiently and accurately, and the detection efficiency can be further improved.

wherein D represents the whole image domain (target region), D1 represents a first region where the difference between a target pixel point x and a non-target pixel point y is larger than zero, D2 represents a second region where the difference between the pixel point x and the non-target pixel point y is smaller than zero, I (-) represents the pixel value of the point, K (-) is a Gaussian kernel function, mid (-) represents the difference between the target pixel point x and the non-target pixel point y, and GP (Gaussian Positive) and GN (Gaussian Negative) are respectively the Positive and Negative parts of the difference between the target pixel point x and the non-target pixel point y, namely the first brightness data and the second brightness data.

(2) Multiplying the values of the two regions by Kernel functions Kernel respectively and adding the values to obtain GP values and GN values respectively;

(3) The value of xn is reassigned to-GN × GP;

The above process can be understood as a gaussian projection process, a gaussian projection radial basis function neural network model is further constructed to achieve the purpose of highlighting the contour, the structure diagram of the model is shown in fig. 2, an input layer is composed of pixel points xn in an image with h pixel points, the difference value of the pixel values of the xn and non-target pixel points is calculated, the result is input to a second layer expansion layer, a third layer is a hidden layer, the hidden layer uses a gaussian function as a hidden function, and the hidden function is:

In another embodiment of the present application, the apparatus further includes a first determining unit, a third obtaining unit, a second determining unit, and a first calculating unit, where the first determining unit is configured to determine a central point after constructing a neural network model, and the central point refers to the input data of the hidden layer of the neural network model; the third acquisition unit is used for acquiring the maximum value and the minimum value of the pixel value of the input data and the number of the pixel points of the input data; the second determining unit is used for determining the variance by adopting an implicit function according to the maximum value of the pixel values, the minimum value of the pixel values and the number of the pixel points; the first calculating unit is configured to calculate an output weight of the neural network model by using a least square method, where the output weight is a weight value of output data of the hidden layer in the calculation result of the output layer. In the scheme, after the neural network model is constructed, the neural network is further trained, so that the trained neural network model is higher in calculation efficiency.

Specifically, the node of the hidden layer is d _i I =1,2, \8230h, h is the total number of input image data, and the corresponding central point is used as each input data, so that the input data can be directly mapped to the hidden layer.

sigma represents variance, xmax represents the maximum value of the pixel value, xmin represents the minimum value of the pixel value, and h represents the number of pixel points, so that a variance mode is directly solved, a more appropriate variance can be obtained for different images, and training efficiency is accelerated to a certain extent.

Since the number of nodes and the variance in the neural network model are determined, the input to the output can be represented by a linear equation, and the output weight can be solved by a least square method, wherein the formula is as follows：

i =1, 2., h, P =1, 2., P, where ω denotes an output weight,

representing the P-th input image, P =1,2, \ 8230, P, P being the total number of input images, d representing the respective center point, i.e. the respective node of the hidden layer.

In another embodiment of the present application, the apparatus further includes a fourth obtaining unit, a second calculating unit, a third determining unit, and a training unit, where the fourth obtaining unit is configured to obtain standard image data after the neural network model is constructed; the second calculating unit is used for calculating an error value between a calculation result output by the neural network model and the standard image data; the third determining unit is used for determining that the neural network model is trained and finished under the condition that the error value is smaller than an error threshold value; the training unit is used for determining that the neural network model needs to be retrained and retraining the neural network model when the error value is larger than or equal to the error threshold value. In this embodiment, the efficiency of the neural network model can be further ensured to be better by detecting the calculated error value of the neural network model and retraining the neural network model under the condition of a larger error value.

In another embodiment of the present application, the detecting unit includes a second determining module, a second obtaining module, and a third determining module, where the second determining module is configured to determine a first edge position and a second edge position of the current image data by using the neural network model; the second acquisition module is used for acquiring a standard center line position of the conveyor belt, wherein the standard center line position refers to a center line position when the conveyor belt is not deviated; and the third determining module is used for determining whether the conveyor belt deviates according to the first edge position, the second edge position and the standard center line position. In this embodiment, whether the conveyor belt has deviated or not can be further accurately determined by counting the first edge position and the second edge position of the processed image and then based on the first edge position, the second edge position, and the standard center line position.

In a specific embodiment of the present application, the third determining module includes a first calculating submodule, a second calculating submodule, a first determining submodule, and a second determining submodule, where the first calculating submodule is configured to calculate a first position difference between the first edge position and the standard centerline position in real time; the second calculating submodule is used for calculating a second position difference value of the second edge position and the standard center line position in real time; the first determining submodule is used for determining that the conveyor belt is off tracking when the first position difference value is smaller than or equal to a position threshold value or the second position difference value is smaller than or equal to the position threshold value; the second determining submodule is used for determining that the conveyor belt is not off tracking under the condition that the first position difference value is larger than the position threshold value and the second position difference value is larger than the position threshold value. In this embodiment, when the distance difference between the first edge position or the second edge position of the conveyor belt and the standard center line position is greater than the position threshold, it indicates that the conveyor belt may not be off tracking at this time, or the off tracking degree is not great, and there is not much influence on the conveyor belt, and if the distance difference between the first edge position or the second edge position and the standard center line position is less than or equal to the position threshold, it indicates that the conveyor belt is off tracking at this time, and the edge of the conveyor belt may approach the standard center line position, so that this embodiment may further accurately determine whether the conveyor belt is off tracking.

Specifically, as shown in fig. 3, the result of the algorithm is shown, the image sensor is installed above the coal preparation plant mine conveyor belt 100, the image pickup direction is along the working direction of the conveyor belt 100, the result of the algorithm can determine the first edge position 200 and the second edge position 300 of the conveyor belt 100, it can be determined that the conveyor belt 100 is working normally, neither the first edge position 200 nor the second edge position 300 reaches the standard center line position 400, and if the first edge position 200 and the second edge position 300 reach the standard center line position 400, a deviation warning is sent.

In another specific embodiment of the present application, the apparatus further includes a generation unit and a control unit, where the generation unit is configured to generate alarm information when it is determined that the conveyor belt has deviated, and the alarm information is used to prompt that the conveyor belt has deviated; the control unit is used for controlling the conveyor belt to pause. In this embodiment, through production alarm information, can in time indicate staff conveyer belt off tracking this moment, and the work of controlling conveyer belt pause again can avoid the conveyer belt off tracking and the dangerous accident that takes place.

And based on the obtained image edge data of each frame, whether the working position of the conveyor belt is normal or not can be quickly judged by detecting the distance between the edge pixel point of the conveyor belt and the standard center line position of the image. Meanwhile, the deviation warning also improves the working stability of the system through multi-frame error accumulation.

The detection device for the deviation of the conveyor belt comprises a processor and a memory, wherein the first acquisition unit, the construction unit, the second acquisition unit, the detection unit and the like are stored in the memory as program units, and the processor executes the program units stored in the memory to realize corresponding functions.

The processor comprises a kernel, and the kernel calls the corresponding program unit from the memory. The kernel can be set to be one or more than one, and the efficiency of detecting whether the conveying belt deviates or not is improved by adjusting the kernel parameters.

The memory may include volatile memory in a computer readable medium, random Access Memory (RAM) and/or nonvolatile memory such as Read Only Memory (ROM) or flash memory (flash RAM), and the memory includes at least one memory chip.

The embodiment of the invention provides a computer-readable storage medium, wherein a program is stored on the computer-readable storage medium, and the program is used for realizing the detection method for the deviation of the conveying belt when being executed by a processor.

The embodiment of the invention provides a processor, which is used for running a program, wherein the program executes the method for detecting the deviation of the conveying belt when running.

An embodiment of the present invention provides an apparatus, where the apparatus includes a processor, a memory, and a program that is stored in the memory and is executable on the processor, and when the processor executes the program, at least the following steps are implemented:

step S103, acquiring current image data of the conveyor belt;

The device herein may be a server, a PC, a PAD, a mobile phone, etc.

The present application further provides a computer program product adapted to perform a program of initializing at least the following method steps when executed on a data processing device:

step S103, acquiring current image data of the conveyor belt;

In order to make the technical solutions of the present application more clearly understood by those skilled in the art, the technical solutions and technical effects of the present application will be described below with reference to specific embodiments.

Examples

The embodiment relates to a method for detecting the deviation of a conveyor belt, as shown in figure 5,

firstly, starting to detect;

the camera sensor collects images, and historical image data are collected at the moment;

acquiring a training sample, and carrying out Gaussian filtering processing on the training sample;

training a Gaussian projected radial basis function neural network model, which comprises the following steps: determining a central point, calculating the variance of an implicit function, and updating a network weight;

determining whether an error value of the neural network model is less than an error threshold;

under the condition that the error value is greater than or equal to the error threshold value, retraining the neural network model;

determining that the neural network model has been trained to end when the error value is less than the error threshold;

the camera sensor collects images, and the current image data are collected at the moment;

obtaining a working sample, and performing Gaussian filtering processing on the working sample;

detecting a current image by adopting a trained Gaussian projection radial basis function neural network model;

determining whether the edge position of the conveyor belt coincides with a standard centerline position;

under the condition of superposition, generating a warning of deviation of the conveyor belt, and controlling the conveyor belt to pause;

determining that the conveyor belt works normally under the condition of misalignment;

the process of detection is ended.

In the above embodiments of the present invention, the descriptions of the respective embodiments have respective emphasis, and for parts that are not described in detail in a certain embodiment, reference may be made to related descriptions of other embodiments.

In the embodiments provided in the present application, it should be understood that the disclosed technology can be implemented in other ways. The above-described embodiments of the apparatus are merely illustrative, and for example, the above-described division of the units may be a logical division, and in actual implementation, there may be another division, for example, multiple units or components may be combined or may be integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed coupling or direct coupling or communication connection between each other may be an indirect coupling or communication connection through some interfaces, units or modules, and may be electrical or in other forms.

The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.

In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.

The integrated unit may be stored in a computer-readable storage medium if it is implemented in the form of a software functional unit and sold or used as a separate product. Based on such understanding, the technical solution of the present invention, which is substantially or partly contributed by the prior art, or all or part of the technical solution may be embodied in a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to perform all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a removable hard disk, a magnetic disk, or an optical disk, and various media capable of storing program codes.

From the above description, it can be seen that the above-mentioned embodiments of the present application achieve the following technical effects:

1) The method for detecting the deviation of the conveyor belt comprises the steps of firstly obtaining historical image data of the conveyor belt, then constructing a neural network model, then obtaining current image data of the conveyor belt, and finally detecting the current image data by adopting the neural network model to determine whether the conveyor belt is deviated. In the scheme, a new neural network model is constructed, the calculation efficiency of the model is higher, the new neural network model is adopted to detect the edges of the conveying belt, the novel neural network model can adapt to a complex environment, the calculation capability is higher than that of a traditional algorithm, the calculation accuracy is higher, the connectivity of the edges of the conveying belt is better, the edges of the conveying belt can be stably detected in an actual environment, whether the conveying belt deviates or not is accurately determined through the detected edges, and therefore the efficiency of detecting whether the conveying belt deviates or not is improved.

2) According to the detection device for the deviation of the conveyor belt, the first acquisition unit acquires historical image data of the conveyor belt, the construction unit constructs the neural network model, the second acquisition unit acquires current image data of the conveyor belt, and the detection unit detects the current image data by adopting the neural network model to determine whether the conveyor belt is deviated. In the scheme, a new neural network model is constructed, the calculation efficiency of the model is higher, the edge of the conveying belt is detected by adopting the new neural network model, the novel neural network model can adapt to a complex environment, the calculation capability is higher than that of a traditional algorithm, the calculation accuracy is higher, the connectivity of the edge of the conveying belt is better, the edge of the conveying belt can be stably detected in an actual environment, whether the conveying belt deviates or not is accurately determined through the detected edge, and therefore the efficiency of detecting whether the conveying belt deviates or not is improved.

The above description is only a preferred embodiment of the present application and is not intended to limit the present application, and various modifications and changes may be made to the present application by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for detecting deviation of a conveyor belt is characterized by comprising the following steps:

acquiring historical image data of a conveyor belt, wherein the historical image data is acquired by image acquisition equipment which is arranged above the conveyor belt;

constructing a neural network model, wherein the neural network is obtained by training a plurality of groups of training data, and each group of training data in the plurality of groups of training data comprises the historical image data and the related information of the conveyor belt corresponding to the historical image data;

acquiring current image data of the conveyor belt;

and detecting the current image data by adopting the neural network model, and determining whether the conveyor belt deviates.

2. The method of claim 1, wherein constructing a neural network model comprises:

acquiring target pixel points and non-target pixel points in a target area, wherein the target pixel points are central pixel points in the target area, the non-target pixel points are pixel points in the target area except the target pixel points, and the target area is an area including the edge of the conveyor belt in a historical image;

determining a first area and a second area, wherein the first area is an area in which the brightness difference value between the target pixel point and the non-target pixel point is greater than or equal to a brightness threshold value, and the second area is an area in which the brightness difference value between the target pixel point and the non-target pixel point is less than the brightness threshold value;

respectively carrying out operation on the image data of the first area and the image data of the second area and a kernel function to obtain first brightness data and second brightness data;

and inputting the first brightness data and the second brightness data into a radial basis function neural network model of Gaussian projection for operation, wherein the radial basis function neural network model of Gaussian projection comprises an input layer, an expansion layer, a hidden layer and an output layer, the input layer is used for receiving input data, the expansion layer is used for performing difference on two adjacent input data, the hidden layer is used for calculating the input data subjected to difference and a Gaussian function, and the output layer is used for outputting a calculation result, wherein the input data is an absolute value of a product of the first brightness data and the second brightness data.

3. The method of claim 2, wherein after constructing the neural network model, the method further comprises:

determining a central point, wherein the central point refers to the input data of the hidden layer of the neural network model;

acquiring the maximum value and the minimum value of the pixel values of the input data and the number of the pixel points of the input data;

determining a variance by adopting an implicit function according to the maximum value of the pixel value, the minimum value of the pixel value and the number of the pixel points;

and calculating an output weight of the neural network model by adopting a least square method, wherein the output weight refers to a weight value occupied by the output data of the hidden layer in the calculation result of the output layer.

4. The method of claim 1, wherein after building the neural network model, the method further comprises:

acquiring standard image data;

calculating an error value between a calculation result output by the neural network model and the standard image data;

determining that the neural network model is trained to be finished when the error value is smaller than an error threshold value;

and under the condition that the error value is greater than or equal to the error threshold value, determining that the neural network model needs to be retrained, and retraining the neural network model.

5. The method of claim 1, wherein detecting the current image data using the neural network model to determine whether the conveyor belt is off-tracking comprises:

determining a first edge position and a second edge position of the current image data by using the neural network model;

acquiring a standard center line position of the conveyor belt, wherein the standard center line position refers to a center line position when the conveyor belt is not deviated;

and determining whether the conveyor belt deviates according to the first edge position, the second edge position and the standard center line position.

6. The method of claim 5, wherein determining whether the conveyor belt is off-track based on the first edge position, the second edge position, and the standard centerline position comprises:

calculating a first position difference between the first edge position and the standard centerline position in real time;

calculating a second position difference value of the second edge position and the standard centerline position in real time;

determining that the conveyor belt has deviated when the first position difference is less than or equal to a position threshold, or the second position difference is less than or equal to the position threshold;

and determining that the conveyor belt is not off tracking under the condition that the first position difference value is greater than the position threshold value and the second position difference value is greater than the position threshold value.

7. The method according to any one of claims 1 to 6, wherein in the event that it is determined that the conveyor belt has deviated, the method further comprises:

generating alarm information, wherein the alarm information is used for prompting that the conveyor belt deviates;

and controlling the conveyor belt to pause.

8. The utility model provides a detection device of conveyer belt off tracking which characterized in that includes:

the device comprises a first acquisition unit, a second acquisition unit and a control unit, wherein the first acquisition unit is used for acquiring historical image data of a conveyor belt, the historical image data is acquired by an image acquisition device, and the image acquisition device is arranged above the conveyor belt;

the building unit is used for building a neural network model, wherein the neural network is obtained by using a plurality of groups of training data through training, and each group of training data in the plurality of groups of training data comprises the historical image data and the related information of the conveyor belt corresponding to the historical image data;

a second acquisition unit configured to acquire current image data of the conveyor belt;

and the detection unit is used for detecting the current image data by adopting the neural network model and determining whether the conveying belt deviates.

9. A computer-readable storage medium, characterized in that the computer-readable storage medium comprises a stored program, wherein the program performs the method of any one of claims 1 to 7.

10. A processor, characterized in that the processor is configured to run a program, wherein the program when running performs the method of any of claims 1 to 7.