Start-stop state monitoring method for mining belt conveyor

Abstract

The invention belongs to the technical field of state monitoring, and particularly relates to a mining belt conveyor start-stop state monitoring method, which comprises the following steps: carrying out linear fitting on representative gray values of sub-regions at the same position in no-load images with the same serial number acquired by two adjacent cameras mounted on a mining belt conveyor, and obtaining a linear fitting result; obtaining an ambient light influence relation and ambient light influence degree of two adjacent cameras in each sub-region, processing a monitoring image collected by the previous camera according to the ambient light influence relation to obtain an ideal image of the next camera, and weighting a difference result of the ideal image and the monitoring image of the next camera according to the ambient light influence degree to obtain a difference result of the monitoring image and the ideal image of the next camera. And the coal flow landslide probability of the monitoring image of the next camera is obtained and is used for judging whether the camera has a fault hidden danger in a corresponding monitoring area on the mining belt conveyor or not, so that the start-stop state of the mining belt conveyor is controlled. According to the invention, the start-stop control accuracy of the mining belt conveyor is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of state monitoring, and more specifically, to a method for monitoring the start and stop states of a mining belt conveyor. Background Art

[0002] The operation of mining belt conveyor is a system work completed by multiple components such as integrated drive device, conveyor belt, conveyor roller, tensioning device, roller and support device. Through the coordinated work of these components, the task of material transportation can be completed stably and efficiently.

[0003] In the related technology, for example, a Chinese patent document with authorization announcement number CN105059868B discloses a belt break detection device and a detection method for a mining belt conveyor. The device includes an image acquisition module, an explosion-proof housing, an ultra-microcomputer module and an intrinsically safe power supply. The image acquisition module is a mining explosion-proof camera, and the shooting direction of the explosion-proof camera is toward the belt tensioning roller. In the patent document, there is no relative movement between the detection device and the tensioning roller. No matter what state the belt conveyor is in, the position of the roller photographed by the camera remains unchanged, which improves the accuracy and reliability of the detection parameters, and only requires one device, avoiding data redundancy and contradictions between multiple detection devices. The detection method in the patent document performs edge detection on the image, compares the coordinates of the belt edge point monitored in real time with the designated normal operating area of ​​the belt conveyor, and determines whether the belt is broken; and filters, compresses and denoises the image, so the judgment accuracy is higher and the judgment speed is faster.

[0004] When monitoring the status of a mining belt conveyor working underground in a coal mine, in order to ensure the clarity of the images captured by the camera, some light sources need to be installed on the top of the coal mine where the mining belt conveyor is installed. The number and distribution of light sources installed near cameras at different locations are different, so the images captured by cameras at different locations are affected differently by the light sources. Even for light sources with the same parameters, their effects will be different as their service life increases, affecting the accuracy of subsequent judgment and control. Summary of the invention

[0005] In order to solve the technical problem that the above-mentioned ambient light affects the accuracy of the state monitoring of the mining belt conveyor, and then affects the accuracy of the start-stop control of the mining belt conveyor, the present invention provides a method for monitoring the start-stop state of the mining belt conveyor, comprising: installing multiple cameras above the mining belt conveyor; respectively, when the mining belt conveyor is unloaded and transporting coal, using each camera to collect multiple unloaded images and multiple monitoring images, and dividing each unloaded image and each monitoring image into multiple sub-areas; for the unloaded images with the same serial number collected by two adjacent cameras, using the representative grayscale values ​​of the sub-areas at the same position in the two unloaded images as the independent variable and the dependent variable, respectively, and performing linear fitting, using the fitting results as the influence relationship of the ambient light of the two adjacent cameras in each sub-area, and according to the fitting, The loss function value of the result is used to determine the degree of ambient light influence of the two adjacent cameras in each sub-area; according to the ambient light influence relationship of the two adjacent cameras in each sub-area, each sub-area in the monitoring image collected by the previous camera of the two adjacent cameras is processed to obtain the ideal image of the next camera of the two adjacent cameras; according to the degree of ambient light influence of the two adjacent cameras in each sub-area, the difference result of each sub-area in the ideal image and the monitoring image of the next camera of the two adjacent cameras is weighted to obtain the coal flow landslide probability of the monitoring image of the next camera; according to the coal flow landslide probability of multiple consecutive monitoring images collected by each camera, it is judged whether there is a hidden fault in the monitoring area corresponding to each camera on the mining belt conveyor, which is used to control the start and stop status of the mining belt conveyor.

[0006] The present invention determines the ambient light influence relationship between two adjacent cameras by comparing two adjacent empty images of each empty area on the belt of a mining belt conveyor under each camera, that is, the empty images with the same serial number captured by two adjacent cameras, and then processes each sub-area in the monitoring image captured by the previous camera of the two adjacent cameras according to the ambient light influence relationship of the two adjacent cameras in each sub-area, so as to obtain an ideal image of the latter camera of the two adjacent cameras. The ideal image is a monitoring image excluding the interference of ambient light. Therefore, the difference result of the ideal image of the camera obtained subsequently and each sub-area in the monitoring image completely reflects the probability of coal flow landslide in the monitoring image of the camera; subsequently, the difference result of the ideal image of the camera obtained subsequently and each sub-area in the monitoring image completely reflects the probability of coal flow landslide in the monitoring image of the camera; The degree of ambient light influence of each camera in each sub-area is taken into account, and the difference results of each sub-area in the camera's ideal image and the monitoring image are weighted. When calculating the probability of coal flow landslide in the camera's monitoring image, the degree of ambient light influence is added, thereby improving the calculation accuracy of the probability of coal flow landslide in the camera's monitoring image; finally, according to the coal flow landslide probability of multiple consecutive monitoring images collected by each camera, when judging whether there is a hidden fault in the monitoring area corresponding to each camera on the mining belt conveyor, the interference of ambient light is accurately eliminated. Therefore, the monitoring accuracy of whether there is a hidden fault in each monitoring area on the mining belt conveyor is improved, thereby improving the accuracy of start and stop control of the mining belt conveyor.

[0007] Preferably, The camera and The cameras are two adjacent cameras. , is the total number of all cameras; In each idle image collected by the camera, The representative grayscale value of the sub-region is taken as the independent variable. In each idle image collected by the camera, The representative grayscale value of the sub-region is used as the independent variable, and then a linear fit is performed. The representative grayscale value of the sub-region refers to the mean grayscale value of all pixels in the sub-region; the expression of the fitting result with the minimum loss function value is used as the first and The camera is in The ambient light influence relationship of each sub-area, and the loss function value is the value of the loss function of the fitting result.

[0008] The present invention determines the expression of the final fitting result by the loss function value, so as to ensure that the expression of the final fitting result can accurately express the first and The camera is in The relationship between the ambient light of each sub-area.

[0009] Preferably, the loss function is equal to the weighted sum of the fitting errors of each data point, including: and The camera captures In the empty image The sub-region is recorded as and , when performing linear fitting, the sub-area The representative gray value and subregions The representative gray value Composition data points; the fitting error of each data point is equal to the absolute value of the difference between the fitting value and the actual value of each data point, and the weighted result of the fitting error of each data point is equal to the product of the fitting error of each data point and the weight of each data point. data points, if or , then The weight of a data point is , otherwise, The weight of a data point is .

[0010] Preferably, determining the degree of ambient light influence of two adjacent cameras in each sub-area based on the loss function value of the fitting result includes: taking the inverse of the loss function value of the fitting result corresponding to the ambient light influence relationship of the two adjacent cameras in each sub-area as the degree of ambient light influence of the two adjacent cameras in each sub-area.

[0011] Preferably, the method of processing each sub-area in the monitoring image captured by the first camera of the two adjacent cameras according to the relationship between the ambient light influences of the two adjacent cameras in each sub-area to obtain an ideal image of the second camera of the two adjacent cameras comprises: The camera and The cameras are adjacent to each other, and the The first camera is the first camera among the two adjacent cameras. The camera is the latter of the two adjacent cameras. , is the total number of all cameras; The camera captures In the monitoring image The gray value of each pixel in the sub-region is used as the independent variable and input into the The camera and In the relationship between the ambient light influence of each sub-area, the area composed of the output dependent variables is taken as the first The camera's The first ideal image An ideal sub-region, , is a fixed quantity.

[0012] The present invention processes each sub-area in the monitoring image captured by the front camera of the two adjacent cameras according to the influence relationship of the ambient light in each sub-area of ​​the two adjacent cameras, and obtains the ideal image of the rear camera of the two adjacent cameras, and the obtained ideal image is a monitoring image excluding the interference of ambient light.

[0013] Preferably, the method of obtaining the coal flow landslide probability of the monitoring image of the next camera includes: The camera and The cameras are adjacent to each other, and the The camera and The cameras are the front camera and the back camera of two adjacent cameras. The camera's Coal flow landslide probability based on monitoring images The calculation formula is: ; , Respectively The camera's Monitoring images and The ideal image In the sub-area The gray value of a pixel, , is a fixed quantity, , For the The camera and The camera is in The degree of influence of ambient light on each sub-area, and are the length and width of the image captured by the camera respectively.

[0014] The present invention obtains the probability of coal flow landslide in the monitoring image of the camera by subtracting the results of each sub-area in the ideal image of the camera and the monitoring image, eliminating the interference of ambient light. Therefore, the subtraction result completely reflects the probability of coal flow landslide in the monitoring image of the camera. The subtraction results of each sub-area in the ideal image of the camera and the monitoring image are weighted by the degree of influence of ambient light in each sub-area of ​​two adjacent cameras. When calculating the probability of coal flow landslide in the monitoring image of the camera, the degree of influence of ambient light is increased, thereby improving the calculation accuracy of the probability of coal flow landslide in the monitoring image of the camera.

[0015] Preferably, the method of judging whether there is a potential fault in the monitoring area corresponding to each camera on the mining belt conveyor according to the coal flow landslide probability of the continuous multiple monitoring images collected by each camera includes: The camera captures monitoring image, if the If the probability of coal flow landslide in the monitoring image is greater than or equal to the preset probability threshold, the The probability of coal flow landslide in the monitoring image is marked as 1. If the probability of coal flow landslide in the monitoring image is less than the preset probability threshold, the The probability of coal flow landslide in the monitoring image is marked as 0; if the number of monitoring images whose probability of coal flow landslide is continuously marked as 1 is greater than the preset number threshold, it means that the first There is a potential malfunction risk in the monitoring area corresponding to each camera on the mining belt conveyor.

[0016] Preferably, the step of dividing each idle image and each monitoring image into a plurality of sub-areas comprises: dividing each idle image into sub-regions, is a fixed number, where the size of each sub-region is equal to , is the size of the image captured by the camera, and are the length and width of the image captured by the camera, respectively, then the size of the unloaded image is equal to , and the length and width of the idle image are equal to and , To round up; divide each monitoring image into sub-regions, is a fixed number, where the size of each sub-region is equal to , is the size of the image captured by the camera, and are the length and width of the image captured by the camera, respectively, then the size of the monitored image is equal to , and the length and width of the monitoring image are equal to and .

[0017] Preferably, the collecting of multiple idle images and multiple monitoring images includes: the time intervals of collecting the idle images and monitoring images by all cameras are the same and equal to , It is equal to the fixed monitoring length divided by the running speed of the mining belt conveyor, and requires that the number of no-load images collected by each camera is the same and equal to , It is equal to the length of the entire belt on the mining belt conveyor divided by the fixed monitoring length; A camera, when The starting time of collecting the idle image and monitoring image is equal to and , , are the time when the mining belt conveyor starts to run without load and the time when it starts to transport coal; When , the starting time of collecting the idle image and monitoring image is equal to and , For the The position of the camera is The distance between the locations of the cameras, It is the running speed of mining belt conveyor.

[0018] The present invention stipulates the starting time for each camera to start collecting no-load images and the time interval for collecting no-load images, so that the no-load areas corresponding to the no-load images with the same serial number collected by all cameras on the belt of the mining belt conveyor are the same; and stipulates the starting time for each camera to start collecting monitoring images and the time interval for collecting monitoring images, so that the coal flow areas corresponding to the monitoring images with the same serial number collected by all cameras on the belt of the mining belt conveyor are the same.

[0019] Preferably, the multiple cameras are installed above the mining belt conveyor, including: fixing a camera in each curved area of ​​the mining belt conveyor, and fixing at least one camera at a spacing in each straight area of ​​the mining belt conveyor; requiring that all cameras are of the same model, and that the installation angles of all cameras can ensure that the viewing angles of the captured coal flow images are the same; and sorting all cameras according to the running track of the mining belt conveyor, and according to the running track of the mining belt conveyor, the serial numbers of all cameras are 1 to 1. , is the total number of all cameras.

[0020] The beneficial effects of the present invention are: The present invention determines the ambient light influence relationship between two adjacent cameras by comparing two adjacent empty images of each empty area on the belt of a mining belt conveyor under each camera, that is, the empty images with the same serial number captured by two adjacent cameras, and then processes each sub-area in the monitoring image captured by the previous camera of the two adjacent cameras according to the ambient light influence relationship of the two adjacent cameras in each sub-area, so as to obtain an ideal image of the latter camera of the two adjacent cameras. The ideal image is a monitoring image excluding the interference of ambient light. Therefore, the difference result of the ideal image of the camera obtained subsequently and each sub-area in the monitoring image completely reflects the probability of coal flow landslide in the monitoring image of the camera; subsequently, the difference result of the ideal image of the camera obtained subsequently and each sub-area in the monitoring image completely reflects the probability of coal flow landslide in the monitoring image of the camera; The degree of ambient light influence of each camera in each sub-area is taken into account, and the difference results of each sub-area in the camera's ideal image and the monitoring image are weighted. When calculating the probability of coal flow landslide in the camera's monitoring image, the degree of ambient light influence is added, thereby improving the calculation accuracy of the probability of coal flow landslide in the camera's monitoring image; finally, according to the coal flow landslide probability of multiple consecutive monitoring images collected by each camera, when judging whether there is a hidden fault in the monitoring area corresponding to each camera on the mining belt conveyor, the interference of ambient light is accurately eliminated. Therefore, the monitoring accuracy of whether there is a hidden fault in each monitoring area on the mining belt conveyor is improved, thereby improving the accuracy of start and stop control of the mining belt conveyor. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The above and other purposes, features and advantages of the exemplary embodiments of the present invention will become readily understood by reading the detailed description below through the accompanying drawings. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein: Figure 1 is a flow chart schematically showing a method for monitoring the start and stop status of a mining belt conveyor in the present invention; Figure 2 is a flowchart schematically showing step S2. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0023] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0024] The embodiment of the present invention discloses a method for monitoring the start and stop status of a mining belt conveyor, referring to Figure 1, including steps S1 to S5: The working principle of mining belt conveyor is very simple. It is mainly composed of driving device, conveyor belt, conveyor roller, tensioning device, roller and supporting device. The operation of mining belt conveyor is a system work completed by multiple components such as driving device, conveyor belt, conveyor roller, tensioning device, roller and supporting device. Through the coordinated work of these components, the task of material transportation can be completed stably and efficiently.

[0025] The working principle of the mining belt conveyor can be simply summarized as follows: the driving device drives the conveyor belt to move, the material is loaded on the conveyor belt, and is transported to the destination through accessories such as conveyor rollers, support devices and rollers; wherein, the driving device usually adopts electric or hydraulic mode, and the main function of the driving device is to provide power for the conveyor belt so that the conveyor belt can move; the conveyor belt is the most core part of the mining belt conveyor, and is the key to transporting materials from one place to another. The material of the conveyor belt is usually rubber, nylon or polyester fiber, etc. The width and length of the conveyor belt can be selected according to actual needs; the conveyor roller is a device that unloads materials from the conveyor belt to the destination. The conveyor roller is usually divided into a head roller and a tail roller, wherein the head roller is driven by a driving device to keep the conveyor belt moving, and the tail roller is an auxiliary device to help the conveyor belt recover; the tensioning device is used to ensure that the conveyor belt always maintains a good tension state; the roller is a device for supporting the conveyor belt, usually composed of bearings and cam shafts, and the number and spacing of the rollers should be determined according to the nature of the material and environmental conditions; the supporting device is mainly used to support the side of the conveyor belt to prevent the conveyor belt from lateral displacement or rupture.

[0026] S1. Install multiple cameras above the mining belt conveyor.

[0027] It should be noted that as the conveyor belt runs, the coal blocks at the initial position will appear in multiple positions later in turn. When comparing and analyzing the images of the same coal flow taken by cameras at different positions, it is possible to determine whether a coal flow landslide has occurred and the location where the coal flow landslide has occurred. Based on whether there are multiple consecutive coal flows that have experienced coal flow landslides at the same position on the mining belt conveyor, it is possible to determine whether there is a fault at a certain position of the mining belt conveyor, and then control the start and stop status of the mining belt conveyor.

[0028] For the mining belt conveyor installed underground in the coal mine for transporting coal, the start and stop status of the mining belt conveyor is controlled by monitoring the status of the process of transporting coal by the mining belt conveyor; therefore, it is necessary to install multiple cameras on the top of the coal mine where the mining belt conveyor is located. The cameras are used to take real-time photos of the coal flow in the fixed viewing angle area on the mining belt conveyor, so as to facilitate the status monitoring of the process of transporting coal by the mining belt conveyor, and the models of all cameras are the same, and the sizes of images captured by all cameras are also the same.

[0029] The design of the mining belt conveyor can be adjusted according to different needs and environmental conditions. In the mining industry, since the space inside the mine may be irregular, the mining belt conveyor may encounter situations where it needs to turn, climb or adapt to terrain changes. Therefore, the mining belt conveyor can be designed to be straight or curved or a combination of straight and curved to adapt to different mine structures and transportation needs, ensuring that it can safely and efficiently complete the coal transportation task; among them, considering that the coal flow is prone to landslides when passing through the curved area of ​​the mining belt conveyor during transportation, therefore, a camera must be fixed in each curved area of ​​the mining belt conveyor, and at least one camera must be fixed at a certain interval in each straight area of ​​the mining belt conveyor; in this embodiment, the spacing is equal to 7 meters. In other embodiments, the specific value of the spacing can be set according to the actual application scenario and needs.

[0030] At the same time, in order to ensure that the coal flow images captured by different cameras can be used for comparative analysis, it is required that the installation angles of all cameras can ensure that the viewing angles of the captured coal flow images are the same.

[0031] Further, all cameras are sorted according to the running track of the mining belt conveyor, and the serial numbers of all cameras are from 1 to , is the total number of all cameras.

[0032] S2. When the mining belt conveyor is empty and transporting coal, each camera collects multiple empty images and multiple monitoring images, and divides each empty image and each monitoring image into multiple sub-areas.

[0033] See the flowchart of step S2. Figure 2 , including steps S201 to S203, specifically: S201. When the mining belt conveyor is unloaded, a plurality of unloaded images are collected by each camera.

[0034] It should be noted that when monitoring the status of a mining belt conveyor working underground in a coal mine, in order to ensure the clarity of the image captured by the camera, some light sources need to be installed on the top of the coal mine where the mining belt conveyor is installed; the number and distribution of light sources installed near cameras at different positions are different, so the images captured by cameras at different positions are affected by different light sources, and even light sources with the same parameters will have different effects as their service life increases; this affects the accuracy of subsequent judgment and control, and therefore, it is necessary to compare the images after removing the influence of the light source to ensure accuracy. The present invention determines the influence of ambient light by comparing the unloaded images taken by the camera.

[0035] Specifically, when the mining belt conveyor is unloaded, multiple unloaded images are collected by each camera. When collecting unloaded images, all light sources and all cameras are required to be turned on. At the same time, there is no coal being transported on the mining belt conveyor, and there are no workers in the space inside the mine. The time intervals for all cameras to collect unloaded images are the same and equal to , then the number of idle images collected by each camera is the same and equal to ;in, It is equal to the fixed monitoring length divided by the running speed of the mining belt conveyor. It is equal to the length of the entire belt on the mining belt conveyor divided by the fixed monitoring length. In this embodiment, the fixed monitoring length is equal to 350 cm. In other embodiments, the specific value of the fixed monitoring length can be set according to the actual application scenario and requirements.

[0036] Among them, for the Cameras, , is the total number of all cameras, The method for determining the starting time for each camera to start collecting idle images is: (1) When , No. The starting time of the camera to collect no-load images is equal to the time when the mining belt conveyor starts to run without load. .

[0037] (2) When At that time, The starting time for each camera to collect idle images is equal to , For the The position of the camera is The distance between the locations of the cameras, is the running speed of the mining belt conveyor; The position of the camera is The distance between the positions of the first and second cameras refers to the distance from the first camera to the second camera along the running track of the mining belt conveyor. The position of the camera moves to the The running distance of the location where the camera is located; To indicate the mining belt conveyor from The position of the camera moves to the The running time of the location of the camera, so for the The first idle image captured by the camera is If the data is collected at the time, then for the Cameras in The first unloaded image collected at the moment is the unloaded area corresponding to the belt of the mining belt conveyor. The moment in the fixed viewing area of ​​a camera is equal to , so the The starting time for each camera to collect idle images is equal to .

[0038] And by stipulating the starting time for each camera to start collecting no-load images and the time interval for collecting no-load images as mentioned above, the no-load areas corresponding to the no-load images with the same serial number collected by all cameras on the belt of the mining belt conveyor are the same, that is, the serial numbers of the no-load images of an no-load area on the belt of the mining belt conveyor under each camera are the same.

[0039] The moment is accurate to seconds (s), so the unit of the time interval is seconds (s); the units of the running distance, fixed monitoring length, and the length of the entire belt on the mining belt conveyor are all centimeters (cm), so the unit of the running speed of the mining belt conveyor is cm / s.

[0040] S202. When the mining belt conveyor transports coal, multiple monitoring images are collected through each camera.

[0041] It should be noted that as the conveyor belt runs, the coal blocks at the initial position will appear in multiple positions later in turn. When comparing and analyzing the images of the same coal flow taken by cameras at different positions, it is possible to determine whether a coal flow landslide has occurred and the location where the coal flow landslide has occurred. Based on whether there are multiple consecutive coal flows that have experienced coal flow landslides at the same position on the mining belt conveyor, it is possible to determine whether there is a fault at a certain position of the mining belt conveyor, and then control the start and stop status of the mining belt conveyor.

[0042] Specifically, when the mining belt conveyor transports coal, multiple monitoring images are collected by each camera. The collected monitoring images are used to monitor the state of the coal flow, thereby realizing fault monitoring of each position of the mining belt conveyor, so as to control the start and stop state of the mining belt conveyor; when collecting monitoring images, it is required to turn on all light sources and all cameras; the time interval for collecting monitoring images by all cameras is the same and equal to , then the number of monitoring images collected by each camera is the same.

[0043] Among them, for the Cameras, , is the total number of all cameras, The method for determining the starting time for each camera to start collecting monitoring images is: (1) When , No. The starting time when the camera starts to collect monitoring images is equal to the time when the mining belt conveyor starts to transport coal. .

[0044] (2) When At that time, The starting time when each camera starts to collect monitoring images is equal to , For the The position of the camera is The distance between the locations of the cameras, is the running speed of the mining belt conveyor; The first monitoring image collected by the camera is If the data is collected at the time, then for the Cameras in The first monitoring image collected at the moment is in the coal flow area corresponding to the belt of the mining belt conveyor. The moment in the fixed viewing area of ​​a camera is equal to , so the The starting time when each camera starts to collect monitoring images is equal to .

[0045] And by stipulating the starting time for each camera to start collecting monitoring images and the time interval for collecting monitoring images as mentioned above, the coal flow area corresponding to the monitoring images with the same serial number collected by all cameras on the belt of the mining belt conveyor is the same, that is, the serial number of the monitoring images of a coal flow area on the belt of the mining belt conveyor under each camera is the same.

[0046] S203: Divide each idle image and each monitoring image into a plurality of sub-areas.

[0047] Specifically, each idle image is divided into sub-regions, is a fixed number, where the size of each sub-region is equal to , is the size of the image captured by the camera, and are the length and width of the image captured by the camera, respectively, then the size of the unloaded image is equal to , and the length and width of the idle image are equal to and , To round up.

[0048] Furthermore, each monitoring image is divided into sub-regions, is a fixed number, where the size of each sub-region is equal to , is the size of the image captured by the camera, and are the length and width of the image captured by the camera, respectively, then the size of the monitored image is equal to , and the length and width of the monitoring image are equal to and .

[0049] In this embodiment, the fixed number equal In other embodiments, a fixed number The specific value can be set according to the actual application scenario and requirements.

[0050] S3. Linear fitting is performed on the representative grayscale values ​​of the sub-regions at the same position in the idle images with the same serial number collected by two adjacent cameras to determine the relationship and degree of ambient light influence of the two adjacent cameras in each sub-region.

[0051] It should be noted that the present invention determines the ambient light influence relationship between two adjacent cameras by comparing two adjacent empty images of an empty area on the belt of the mining belt conveyor under each camera, that is, the empty images with the same serial number captured by two adjacent cameras.

[0052] Specifically, for idle images with the same serial number captured by two adjacent cameras, the representative grayscale values ​​of the sub-areas at the same position in the two idle images are respectively used as independent variables and dependent variables and linear fitting is performed, and the fitting results are used as the ambient light influence relationship of the two adjacent cameras in each sub-area, and the loss function value is the value of the loss function of the fitting result; and according to the loss function value of the fitting result, the degree of ambient light influence of the two adjacent cameras in each sub-area is determined.

[0053] Among them, The camera and The cameras are adjacent to each other, and the The first camera is the first camera among the two adjacent cameras. The camera is the latter of the two adjacent cameras. , is the total number of all cameras; The camera captures The empty image is denoted as , will The camera captures The empty image is denoted as , is the serial number of the idle image collected by the camera, , the number of all idle images captured by the camera is equal to ; will load the image Middle The sub-region is recorded as , the empty image Middle The sub-region is recorded as , is the sub-region number, , is a fixed number, and the number of all sub-regions in each empty image is equal to .

[0054] The first In each idle image collected by the camera, The representative grayscale value of the sub-region is taken as the independent variable. In each idle image collected by the camera, The representative grayscale value of the sub-region is used as the independent variable, and then a linear fit is performed. The representative grayscale value of the sub-region refers to the mean grayscale value of all pixels in the sub-region; the expression of the fitting result with the minimum loss function value is used as the first and The camera is in The relationship between the ambient light of each sub-area.

[0055] Among them, when performing linear fitting, the sub-area Representative grayscale values ​​and sub-regions The representative gray value of data points, sub-areas The representative gray value is the independent variable, and the sub-area The representative gray value of is the dependent variable, then the loss function is equal to the sum of the weighted results of the fitting errors of each data point, the fitting error of each data point is equal to the absolute value of the difference between the fitting value and the actual value of each data point, and the weighted result of the fitting error of each data point is equal to the product of the fitting error of each data point and the weight of each data point. data points, if or , then The weight of a data point is , otherwise, The weight of a data point is .

[0056] in, , Sub-areas Representative grayscale values ​​and sub-regions The representative gray value of To obtain the minimum function, is the maximum value function.

[0057] It should be noted that for data points whose grayscale values ​​of the independent variable or dependent variable are equal to 0 or 255, they cannot accurately express the ambient light influence relationship. Therefore, their weights are set equal to 0 to avoid their influence on the accuracy of the fitting results, and further avoid their influence on the accuracy of the ambient light influence relationship between two adjacent cameras in each sub-area.

[0058] Further, determining the degree of ambient light influence of two adjacent cameras in each sub-area according to the loss function value of the fitting result includes: taking the inverse of the loss function value of the fitting result corresponding to the ambient light influence relationship of the two adjacent cameras in each sub-area as the degree of ambient light influence of the two adjacent cameras in each sub-area.

[0059] It should be noted that the smaller the loss function value of the fitting result is, the more accurate the obtained relationship between the ambient light influence of the two adjacent cameras in each sub-area is. Accordingly, when the monitoring image collected by the first camera of the two adjacent cameras is processed according to the ambient light influence relationship of the two adjacent cameras in each sub-area, the ideal image of the second camera of the two adjacent cameras can be obtained more accurately, and the degree of ambient light influence of the two adjacent cameras in each sub-area is smaller.

[0060] S4. According to the relationship between the ambient light influences of two adjacent cameras in each sub-area, each sub-area in the monitoring image collected by the first camera of the two adjacent cameras is processed to obtain an ideal image of the second camera.

[0061] Specifically, according to the relationship of the ambient light influence of two adjacent cameras in each sub-area, each sub-area in the monitoring image collected by the first camera of the two adjacent cameras is processed to obtain an ideal image of the second camera of the two adjacent cameras.

[0062] Among them, The camera and The cameras are adjacent to each other, and the The first camera is the first camera among the two adjacent cameras. The camera is the latter of the two adjacent cameras. , is the total number of all cameras; The camera captures In the monitoring image The gray value of each pixel in the sub-region is used as the independent variable and input into the The camera and In the relationship between the ambient light influence of each sub-area, the area composed of the output dependent variables is taken as the first The camera's The first ideal image An ideal sub-region, , is a fixed quantity.

[0063] In particular, since both the independent variable and the dependent variable in the ambient light influence relationship belong to grayscale values, the grayscale value range is Therefore, when the output dependent variable is less than 0, the output dependent variable is corrected to 0, and when the output dependent variable is greater than 255, the output dependent variable is corrected to 255.

[0064] It should be noted that, by utilizing the obtained ambient light influence relationship, each sub-area in the monitoring image collected by the first camera of two adjacent cameras is processed to obtain the ideal image of the second camera of the two adjacent cameras, so as to facilitate the subsequent difference between the ideal image of the camera and the sub-area in the monitoring image to obtain whether there are hidden fault hazards in the monitoring area corresponding to the camera on the mining belt conveyor, accurately eliminate the interference of ambient light, and improve the monitoring accuracy of whether there are hidden fault hazards in each monitoring area on the mining belt conveyor.

[0065] S5. According to the degree of influence of ambient light, the difference results of each sub-area in the ideal image of the rear camera and the monitoring image are weighted to obtain the coal flow landslide probability of the monitoring image of the rear camera. According to the coal flow landslide probability of multiple consecutive monitoring images collected by each camera, it is judged whether there is a hidden danger of fault in the monitoring area corresponding to each camera on the mining belt conveyor, which is used to control the start and stop status of the mining belt conveyor.

[0066] Specifically, according to the degree of influence of ambient light in each sub-area of ​​two adjacent cameras, the difference results of the ideal image of the rear camera and each sub-area in the monitoring image are weighted to obtain the probability of coal flow landslide in the monitoring image of the rear camera; according to the coal flow landslide probability of multiple consecutive monitoring images collected by each camera, it is judged whether there is a hidden fault hazard in the monitoring area corresponding to each camera on the mining belt conveyor, which is used to control the start and stop status of the mining belt conveyor.

[0067] Among them, The camera and The cameras are adjacent to each other, and the The first camera is the first camera among the two adjacent cameras. The first camera is the latter of the two adjacent cameras. The camera and The degree of influence of ambient light on each sub-area of ​​the camera The camera's The ideal image and The difference results of each sub-region in the monitoring image are weighted to obtain the The camera's The probability of coal flow landslide based on the monitoring image, , is a fixed number; then The camera's The calculation formula of the coal flow landslide probability of the monitoring image is: ; In the formula, For the The camera's The probability of coal flow landslide based on the monitoring image, For the The camera's The monitoring image In the sub-area The gray value of a pixel, For the The camera's The ideal image In the ideal sub-region The gray value of a pixel, , No. The camera's The number of all sub-regions in the monitoring image and the The camera's The number of all sub-regions in an ideal image is equal to , is a fixed quantity, is the size of the image captured by the camera, and are the length and width of the image captured by the camera, , No. The size of the sub-regions and The size of the ideal sub-region is equal to , then The number of all pixels in the sub-region and the The number of all pixels in an ideal sub-region is equal to , To take the absolute value, For the The camera and The camera is in The degree of influence of ambient light on each sub-area, Indicates The camera's The ideal image and In the monitoring image The difference result of each sub-region.

[0068] It should be noted that the probability of coal flow landslide in the camera's monitoring image is obtained by subtracting the results of each sub-area in the camera's ideal image and the monitoring image, eliminating the interference of ambient light. Therefore, the subtraction result fully reflects the probability of coal flow landslide in the camera's monitoring image; the subtraction results of the camera's ideal image and each sub-area in the monitoring image are weighted by the degree of ambient light influence of two adjacent cameras in each sub-area. When calculating the probability of coal flow landslide in the camera's monitoring image, the degree of ambient light influence is increased, thereby improving the calculation accuracy of the probability of coal flow landslide in the camera's monitoring image.

[0069] Among them, for the The camera captures monitoring image, if the If the probability of coal flow landslide in the monitoring image is greater than or equal to the preset probability threshold, the The probability of coal flow landslide in the monitoring image is marked as 1. If the probability of coal flow landslide in the monitoring image is less than the preset probability threshold, the The probability of coal flow landslide in the monitoring image is marked as 0; if the number of monitoring images whose probability of coal flow landslide is continuously marked as 1 is greater than the preset number threshold, it means that the first If there is a hidden danger of failure in the monitoring area corresponding to a camera on the mining belt conveyor, the mining belt conveyor is controlled to stop so that maintenance personnel can inspect and repair it. The failure of the monitoring area corresponding to a camera on the mining belt conveyor.

[0070] In this embodiment, the preset probability threshold is equal to In other embodiments, the specific value of the preset probability threshold can be set according to the actual application scenario and requirements, and the value range is .

[0071] In this embodiment, the preset quantity threshold is equal to In other embodiments, the specific value of the preset probability threshold can be set according to the actual application scenario and requirements, and the value range is .

[0072] It should be noted that, when finally judging whether there are hidden fault hazards in the monitoring area corresponding to each camera on the mine belt conveyor based on the coal flow landslide probability of multiple consecutive monitoring images collected by each camera, the interference of ambient light is accurately eliminated. Therefore, the monitoring accuracy of whether there are hidden fault hazards in each monitoring area on the mine belt conveyor is improved, thereby improving the accuracy of start and stop control of the mine belt conveyor.

[0073] In the description of this specification, "plurality" or "several" means at least two, such as two, three or more, etc., unless otherwise clearly and specifically defined.

[0074] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, alternatives to the embodiments of the present invention described herein may be employed.

Claims

1. A method for monitoring the start and stop status of a mining belt conveyor, characterized in that: include: Install multiple cameras above the mining belt conveyor; When the mining belt conveyor is empty and transporting coal, each camera collects multiple empty images and multiple monitoring images, and divides each empty image and each monitoring image into multiple sub-areas; For the idle images with the same serial number collected by two adjacent cameras, the representative grayscale values ​​of the sub-regions at the same position in the two idle images are used as the independent variable and the dependent variable, respectively, and linear fitting is performed. The fitting results are used as the relationship of the ambient light influence of the two adjacent cameras in each sub-region, and the degree of ambient light influence of the two adjacent cameras in each sub-region is determined according to the loss function value of the fitting results; According to the relationship between the ambient light influences of two adjacent cameras in each sub-area, each sub-area in the monitoring image collected by the first camera of the two adjacent cameras is processed to obtain an ideal image of the second camera of the two adjacent cameras; According to the degree of influence of ambient light in each sub-area of ​​two adjacent cameras, the difference results of each sub-area in the ideal image of the latter camera and the monitoring image of the latter camera are weighted to obtain the probability of coal flow landslide in the monitoring image of the latter camera; According to the coal flow landslide probability of multiple continuous monitoring images collected by each camera, it is judged whether there is a hidden fault hazard in the monitoring area corresponding to each camera on the mining belt conveyor, which is used to control the start and stop status of the mining belt conveyor.

2. The method for monitoring the start and stop status of a mining belt conveyor according to claim 1, characterized in that: No. The camera and The cameras are two adjacent cameras. , is the total number of all cameras; In each idle image collected by the camera, The representative grayscale value of the sub-region is taken as the independent variable. In each idle image collected by the camera, The representative grayscale value of each sub-region is used as an independent variable, and then a linear fit is performed. The representative grayscale value of the sub-region refers to the mean grayscale value of all pixels in the sub-region; The expression of the fitting result with the minimum loss function value is used as the first and The camera is in The ambient light influence relationship of each sub-area, and the loss function value is the value of the loss function of the fitting result.

3. The method for monitoring the start and stop status of a mining belt conveyor according to claim 2, characterized in that: The loss function is equal to the weighted sum of the fitting errors of each data point, including: Separately and The camera captures In the empty image The sub-region is recorded as and , when performing linear fitting, the sub-area The representative gray value and subregions The representative gray value Composition data points; The fitting error of each data point is equal to the absolute value of the difference between the fitting value and the actual value of each data point, and the weighted result of the fitting error of each data point is equal to the product of the fitting error of each data point and the weight of each data point. data points, if or , then The weight of a data point is , otherwise, The weight of a data point is .

4. The method for monitoring the start and stop status of a mining belt conveyor according to claim 1, characterized in that: The determining, based on the loss function value of the fitting result, the degree of ambient light influence of two adjacent cameras in each sub-area includes: taking the inverse of the loss function value of the fitting result corresponding to the ambient light influence relationship of the two adjacent cameras in each sub-area as the degree of ambient light influence of the two adjacent cameras in each sub-area.

5. The method for monitoring the start and stop status of a mining belt conveyor according to claim 1, characterized in that: The method of processing each sub-region in the monitoring image collected by the first camera of the two adjacent cameras according to the relationship of the ambient light influence in each sub-region of the two adjacent cameras to obtain an ideal image of the second camera of the two adjacent cameras includes: No. The camera and The cameras are adjacent to each other, and the The first camera is the first camera among the two adjacent cameras. The camera is the latter of the two adjacent cameras. , is the total number of all cameras; The camera captures In the monitoring image The gray value of each pixel in the sub-region is used as the independent variable and input into the The camera and In the relationship between the ambient light influence of each sub-area, the area composed of the output dependent variables is taken as the first The camera's The first ideal image An ideal sub-region, , is a fixed quantity.

6. The method for monitoring the start and stop status of a mining belt conveyor according to claim 1, characterized in that: The method of obtaining the coal flow landslide probability of the monitoring image of the latter camera includes: No. The camera and The cameras are adjacent to each other, and the The camera and The first and second cameras are the first and second cameras of two adjacent cameras. The camera's Coal flow landslide probability based on monitoring images The calculation formula is: ； , Respectively The camera's Monitoring images and The ideal image In the sub-area The gray value of a pixel, , is a fixed quantity, , For the The camera and The camera is in The degree of influence of ambient light on each sub-area, and are the length and width of the image captured by the camera respectively.

7. The method for monitoring the start and stop status of a mining belt conveyor according to claim 1, characterized in that: The method of judging whether there is a potential fault in the monitoring area corresponding to each camera on the mining belt conveyor according to the coal flow landslide probability of the continuous multiple monitoring images collected by each camera includes: For The camera captures monitoring image, if the If the probability of coal flow landslide in the monitoring image is greater than or equal to the preset probability threshold, the The probability of coal flow landslide in the monitoring image is marked as 1. If the probability of coal flow landslide in the monitoring image is less than the preset probability threshold, the The probability of coal flow landslide in the monitoring image is marked as 0; if the number of monitoring images whose probability of coal flow landslide is continuously marked as 1 is greater than the preset number threshold, it means that the first There is a potential malfunction risk in the monitoring area corresponding to each camera on the mining belt conveyor.

8. The method for monitoring the start and stop status of a mining belt conveyor according to claim 1, characterized in that: The step of dividing each idle image and each monitoring image into a plurality of sub-areas includes: Divide each idle image into sub-regions, is a fixed number, where the size of each sub-region is equal to , is the size of the image captured by the camera, and are the length and width of the image captured by the camera, respectively, then the size of the unloaded image is equal to , and the length and width of the idle image are equal to and , To round up; Each monitoring image is divided into sub-regions, is a fixed number, where the size of each sub-region is equal to , is the size of the image captured by the camera, and are the length and width of the image captured by the camera, respectively, then the size of the monitored image is equal to , and the length and width of the monitoring image are equal to and .

9. The method for monitoring the start and stop status of a mining belt conveyor according to claim 1, characterized in that: The collecting of multiple idle images and multiple monitoring images includes: The time intervals for all cameras to collect idle images and monitoring images are the same and equal to , It is equal to the fixed monitoring length divided by the running speed of the mining belt conveyor, and requires that the number of no-load images collected by each camera is the same and equal to , It is equal to the length of the entire belt on the mining belt conveyor divided by the fixed monitoring length; A camera, when The starting time of collecting the idle image and monitoring image is equal to and , , are the time when the mining belt conveyor starts to run without load and the time when it starts to transport coal; When , the starting time of collecting the idle image and monitoring image is equal to and , For the The position of the camera is The distance between the locations of the cameras, It is the running speed of mining belt conveyor.

10. The method for monitoring the start and stop status of a mining belt conveyor according to claim 1, characterized in that: The method comprises installing a plurality of cameras above the mining belt conveyor, comprising: A camera must be fixed in each curved area of ​​the mining belt conveyor, and at least one camera must be fixed in each straight area of ​​the mining belt conveyor according to the spacing; All cameras must be of the same model and installed at the same angle to ensure that the coal flow images captured have the same viewing angle; All cameras are sorted according to the running track of the mining belt conveyor, and the serial numbers of all cameras are from 1 to , is the total number of all cameras.

Images