Image recognition method of scraper based on image processing

Abstract

The present invention relates to the field of remote control technology, and in particular to a scraper machine image recognition method based on image processing; the method is implemented by a data acquisition module, a camera module, a color feature analysis module, a texture feature analysis module, a contour analysis module, an image recognition output module and an intelligent management module; the color, texture and contour feature analysis of the scraper machine image are respectively performed by the color feature analysis module, the texture feature analysis module and the contour analysis module to output the scraper machine image recognition result, which can more comprehensively analyze the scraper machine image and improve the accuracy of the scraper machine image recognition result; the scraper machine is stopped by controlling the distance between the responsible employee and the scraper, which can avoid the responsible employee stopping the scraper machine in advance before reaching the position of the abnormal warning scraper, resulting in unreasonable downtime and further waste of production resources of the coal preparation plant; and the abnormal scraper machine can be repaired while maximally ensuring the working efficiency of the coal preparation plant.

Description

Technical Field

[0001] The present invention relates to the technical field of image recognition, and in particular to a scraper image recognition method based on image processing. Background Art

[0002] Scrapers are transport machines used in coal mining working faces. They are primarily used for transporting coal in mining working faces and along mining areas. They can also be used in coal lanes and semi-coal rock lanes. They can transport upwards or downwards. Their transport capacity is unaffected by the size and humidity of the cargo. Their low height facilitates loading, their length is easily adjustable, and their robustness allows them to be used on working faces where coal is being broken and loaded. Scrapers play a key role in the coal preparation process in coal mines. Therefore, identifying and monitoring the status of scrapers is extremely important.

[0003] The working environment of the scraper is complex, and the image will be affected by environmental factors, resulting in difficulty in recognition, which makes it impossible to accurately detect the status and abnormal conditions of the scraper, and thus causes failure and damage to the scraper. Summary of the Invention

[0004] To this end, the present invention provides a scraper image recognition method based on image processing to overcome the problems mentioned in the above background technology.

[0005] To achieve the above object, the present invention provides a scraper machine image recognition method based on image processing, comprising the following steps:

[0006] S1: Collect scraper machine image information, and pre-store standard scraper machine images;

[0007] S2: Analyze the scraper to obtain the monitoring interval and risk level, where the risk level includes high-risk scraper, medium-risk scraper and low-risk scraper;

[0008] S3: Monitor the scraper according to the monitoring interval;

[0009] S4: Extract the scraper machine image and perform color feature extraction on the scraper machine image to obtain a scraper machine color distribution map. Perform color feature analysis based on the scraper machine color distribution map to obtain a color difference value. The specific steps are as follows:

[0010] S41: Identify the scraper color distribution map to obtain the RGB value corresponding to each part in the image, and compare and analyze the RGB values of the same part in the scraper images at adjacent acquisition moments to obtain the difference parts;

[0011] S42: Randomly select the scraper machine color distribution map corresponding to a certain sampling time, and perform S41 on it together with the scraper machine color distribution maps corresponding to the two adjacent sampling times to obtain the front difference part and the number of the front difference parts, the rear difference part and the number of the rear difference parts, and record the number of the front difference parts and the number of the rear difference parts as r1 and r2, respectively;

[0012] S43: randomly taking the RGB value of a certain part in the scraper color distribution map corresponding to a certain acquisition moment, and performing difference calculation on the RGB value of the part and the adjacent part to obtain an adjacent difference. When the adjacent difference is greater than the set adjacent part difference, the part is recorded as an adjacent difference. Similarly, the number of adjacent differences of each part in the scraper color distribution map is obtained; the number of adjacent differences of each part is compared and analyzed with the set number interval to obtain high-difference parts, medium-difference parts and slight-difference parts; counting the number of high-difference parts, medium-difference parts and slight-difference parts in the scraper color distribution map, and recording them as r3, r4 and r5 respectively;

[0013] S44: r1, r2, r3, r4 and r5 are set by the formula The color difference value RZ is calculated, where b1, b2, b3, b4 and b5 are the set proportional coefficients respectively;

[0014] S5: extracting the scraper image and performing texture feature extraction on the image to obtain texture feature information, and performing texture analysis on the texture feature information to obtain a texture difference value;

[0015] S6: extracting the scraper machine image and the standard scraper machine image and performing contour feature extraction on them to obtain the scraper machine contour and the standard scraper machine contour, and performing contour analysis on the two to obtain a contour difference value;

[0016] S7: Comprehensively analyzing the risk level, color difference value, texture difference value, and contour difference value of the scraper to output an image recognition result of the scraper;

[0017] S8: The scraper performs a corresponding pre-execution strategy based on the scraper image recognition result.

[0018] Furthermore, the scraper is analyzed to obtain the monitoring interval and risk level. The specific steps are as follows:

[0019] 201: Obtain the working time and the service time of the scraper, and record them as q1 and q2 respectively;

[0020] 202: Obtain the number of maintenance times of the scraper, the maintenance time and the maintenance duration of each maintenance, use the maintenance time as the horizontal axis, draw vertical bars of equal width at the corresponding maintenance time, and the length of the vertical bars of equal width is equal to the corresponding maintenance duration in value to create a scraper maintenance report;

[0021] 203: Calculate the difference between the maintenance times corresponding to two adjacent maintenance times to obtain the maintenance interval duration, and compare the interval maintenance duration with the set duration interval to obtain the long interval, the medium interval, and the short interval. Count the number of long intervals, medium intervals, and short intervals, and record them as q3, q4, and q5, respectively.

[0022] 204: Calculate the slope of the line segment connecting the midpoints of two adjacent equal-width strips. When the slope is greater than zero, record it as an increasing slope; when the slope is less than zero, record it as a decreasing slope. When any two increasing slopes have overlapping maintenance moments, the two increasing slopes are recorded as a group of adjacent increasing slopes. The number of groups of adjacent increasing slopes is counted and recorded as q6. When any two decreasing slopes have overlapping maintenance moments, the two decreasing slopes are recorded as a group of adjacent decreasing slopes. The number of groups of adjacent decreasing slopes is counted and recorded as q7.

[0023] 205: q1, q2, q3, q4, q5, q6 and q7 are set by the formula The maintenance usage index QZ is calculated, where a1, a2 and a3 are the set proportional coefficients, and the values of a1, a2 and a3 are all natural numbers greater than zero; the maintenance usage index is compared and analyzed with the set index range to obtain high-risk scraper, medium-risk scraper and low-risk scraper.

[0024] Furthermore, texture analysis is performed on the texture feature information to obtain a texture difference value. The specific steps are as follows:

[0025] Extracting a scraper machine image and a standard scraper machine image, and performing texture feature extraction on the scraper machine image and the standard scraper machine image using a local binary pattern to obtain texture feature information and standard texture feature information, wherein the texture feature information includes texture energy, texture contrast, and texture entropy; and the standard texture feature information includes standard texture energy, standard texture contrast, and standard texture entropy;

[0026] The texture energy, texture contrast and texture entropy are respectively calculated by difference with the standard texture energy, standard texture contrast and standard texture entropy to obtain energy difference, contrast difference and entropy difference, and they are recorded as l1, l2 and l3 respectively; l1, l2 and l3 are substituted into the set formula LZ = c1×l1+c2×l2+c3×l3 to calculate the texture contrast value LZ, where c1, c2 and c3 are the set proportional coefficients respectively.

[0027] Furthermore, the scraper machine image and the standard scraper machine image are subjected to contour analysis to obtain a contour difference value. The specific steps are as follows:

[0028] The scraper machine image and the standard scraper machine image are subjected to contour extraction using an edge detection algorithm to obtain the scraper machine contour and the standard scraper machine contour, and the contour area and the standard contour area are calculated;

[0029] The scraper profile is overlapped and compared with the standard scraper profile to obtain the overlap area. When the overlap area ≠ profile area ≠ standard profile area, the difference between the standard profile area and the overlap area is calculated to obtain the profile difference value, which is recorded as LZ.

[0030] Furthermore, the specific steps for outputting the image recognition results of the scraper are as follows:

[0031] Set different risk levels to correspond to a risk coefficient, match the risk level of the scraper with all the set risk levels to obtain the corresponding risk coefficient, and record it as FZ;

[0032] The risk factor FZ, color difference value RZ, texture comparison value WZ and contour difference value LZ are calculated by the set formula FRZ = d1 × FZ + d2 × RZ + d3 × WZ + d4 × LZ to obtain the state value FRZ, where d1, d2, d3 and d4 are the set proportional coefficients respectively;

[0033] The state value is compared and analyzed with the set state interval to generate a fault warning and record the scraper corresponding to the fault warning as a faulty scraper; the abnormal warning and record the scraper corresponding to the abnormal warning as an abnormal scraper; the fault warning signal and the faulty scraper, the abnormal warning signal and the abnormal scraper are recorded as the image recognition results of the scraper.

[0034] Furthermore, the scraper performs a corresponding pre-execution strategy based on the scraper image recognition result. The specific steps are as follows:

[0035] When the scraper machine image recognition result is a fault warning, the number and location of the faulty scraper machine are retrieved, and the corresponding responsible employee and the location of the responsible employee are retrieved based on the number of the faulty scraper machine; the target scraper machine and the target path are selected from the spare scrapers, and the number and location of the faulty scraper machine, the number and location of the target scraper machine, and the target path are sent to the responsible employee's mobile terminal;

[0036] When the scraper machine image recognition result is an abnormality warning, the number and location of the abnormal scraper machine will be retrieved, and the corresponding responsible employee and the location of the responsible employee will be retrieved based on the number of the abnormal scraper machine; the distance difference between the location of the responsible employee and the location of the scraper machine will be calculated to obtain the distance;

[0037] Retrieve the responsible employee's historical maintenance times, as well as the historical distance and duration of each maintenance, and analyze them to obtain the historical average speed;

[0038] Divide the distance between the routes by the historical average speed to obtain the estimated duration, and then control the scraper to continue working for a time equal to the estimated duration before stopping.

[0039] Beneficial effects of the present invention:

[0040] 1. The color feature analysis module, texture feature analysis module and contour analysis module are used to analyze the color, texture and contour features of the scraper image respectively to output the scraper image recognition results, which can more comprehensively analyze the scraper image and improve the accuracy of the scraper image recognition results.

[0041] 2. Through the intelligent management module, the strategy of executing faulty scrapers and abnormal scrapers is implemented, so that faulty scrapers and spare scrapers can be replaced in the shortest time to ensure the working efficiency and safety of the scrapers. At the same time, the scraper shutdown control is carried out according to the distance between the responsible employee and the abnormal scraper, so as to avoid the responsible employee stopping the scraper in advance before reaching the location of the abnormal warning scraper, resulting in unreasonable downtime and waste of production resources of the coal preparation plant. The abnormal scraper can be repaired while maximizing the working efficiency of the coal preparation plant. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 It is a schematic flow diagram of the present invention;

[0043] Figure 2 It is a schematic diagram of module connection of the present invention;

[0044] Figure 3 It is a schematic diagram of a scraper maintenance report of the present invention. DETAILED DESCRIPTION

[0045] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are merely used to explain the present invention and are not intended to limit the present invention.

[0046] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0047] like Figure 1-2 As shown, the present invention is a scraper image recognition method based on image processing, which is implemented by a data acquisition module, a camera module, a color feature analysis module, a texture feature analysis module, a contour analysis module, an image recognition output module and an intelligent management module, and includes the following steps:

[0048] S1: The data acquisition module collects scraper information and scraper image information and sends it to the image database for storage. The scraper image information includes the scraper image and the acquisition time corresponding to the image. The scraper information includes the number of maintenance times, the maintenance time and the maintenance duration of each maintenance.

[0049] S2: The camera module analyzes the scraper information to obtain the monitoring interval duration and sends the monitoring interval duration to the image database module. The specific steps of the analysis are as follows:

[0050] S21: Obtain the working time and the service time of the scraper. It should be noted that the working time refers to the most recent startup time of the scraper from the current time, and the time difference between the startup time and the current time is calculated to obtain the working time; the service time refers to the time between the scraper being put into use and the current time, and the working time and service time are recorded as q1 and q2 respectively.

[0051] S22: Figure 3 As shown, the number of maintenance times of the scraper and the maintenance time and maintenance duration of each maintenance are extracted. With the maintenance time as the horizontal axis, equal-width vertical bars are drawn at the corresponding maintenance time. The length of the equal-width vertical bars is equal to the corresponding maintenance duration in value to establish a scraper maintenance report;

[0052] S23: Calculate the difference between the maintenance times corresponding to two adjacent maintenance times to obtain a maintenance interval, and compare the maintenance interval with the set interval. When the maintenance interval is greater than the maximum value in the set interval, it means that the interval between the two adjacent maintenances is relatively long, and the maintenance interval is recorded as a long interval; when the maintenance interval is within the set interval, it is recorded as a medium interval; when the maintenance interval is less than the minimum value in the set interval, it is recorded as a short interval; count the number of long intervals, medium intervals, and short intervals, and record them as q3, q4, and q5, respectively.

[0053] S24: Use the least squares method to calculate the slope of the line segment connecting the midpoints of two adjacent equal-width strips, and record them as Where i = 1, 2, 3...n1, i represents the maintenance time, n1 represents the total number of maintenance times; It represents the slope of the line segment composed of two adjacent maintenance moments with maintenance moments i and i+1; all slopes are compared and correlated. When the slope is greater than zero, it is recorded as an increasing slope, and when the slope is less than zero, it is recorded as a decreasing slope; when there is an overlapping maintenance moment between any two increasing slopes, the two increasing slopes are recorded as a group of adjacent increasing slopes; specifically, there is and The increasing slope of and There is a coincidence maintenance time of 3, and There is a coincidence maintenance time of 6, then and and Record them as two groups of adjacent increasing slopes respectively; count the number of groups of adjacent increasing slopes and record it as q6;

[0054] When there is a coincident maintenance moment between any two descending slopes, the two descending slopes are recorded as a group of adjacent descending slopes, specifically: The increasing slope of and There is a coincidence maintenance time of 9, and There are 10 overlapping maintenance times. and There is a coincidence maintenance time of 11, then and and and They are recorded as three groups of adjacent descending slopes; the number of groups of adjacent descending slopes is counted and recorded as q7;

[0055] S25: q1, q2, q3, q4, q5, q6 and q7 are calculated by the formula The maintenance usage index QZ is calculated, where a1, a2 and a3 are respectively set proportional coefficients, and the values of a1, a2 and a3 are all natural numbers greater than zero; it can be seen from the formula that the longer the working time and the longer the usage time, the greater the maintenance usage index; the greater the number of medium intervals and short intervals, the greater the maintenance usage index; the greater the number of groups of adjacent increasing slopes and the smaller the number of groups of adjacent decreasing slopes, the greater the maintenance usage index; the larger the maintenance usage index, the greater the degree of aging and deterioration of the scraper, and the greater the potential risk of failure or damage; the maintenance usage index is compared and analyzed with the set index interval. When the maintenance usage index is greater than the maximum value in the set index interval, the scraper is recorded as a high-risk scraper; when the maintenance usage index is within the set index interval, the scraper is recorded as a medium-risk scraper; when the maintenance usage index is less than the minimum value in the set index interval, the scraper is recorded as a low-risk scraper;

[0056] S26: Multiplying the maintenance usage index by the corresponding monitoring conversion coefficient to obtain a monitoring interval, and sending the monitoring interval and risk level of the scraper to the image database;

[0057] S3: The image database module obtains the acquisition time of the last scraper machine image information. When the time difference between the acquisition time and the current system time is equal to the monitoring interval, the image acquisition instruction is sent to the camera module; the image database pre-stores the standard scraper machine image;

[0058] When the camera module receives an image acquisition instruction, it controls the camera angle and focal length to align with the scraper to monitor the scraper;

[0059] S4: The color feature analysis module extracts the scraper image from the image database, and performs color feature extraction on the scraper image to obtain a scraper color distribution map. The color feature analysis is performed based on the scraper color distribution map. The specific steps are as follows:

[0060] S41: using a photo identifier to identify the scraper machine color distribution map to obtain the RGB value corresponding to each part in the image, and comparing and analyzing the RGB values of the same part in the scraper machine images at adjacent acquisition moments. When the difference in the RGB value of the same part is greater than a set same-part difference, it indicates that the RGB value of the part in the scraper machine color distribution map corresponding to the adjacent acquisition moments is greatly different, and the part is recorded as a difference part;

[0061] S42: randomly select the scraper machine color distribution map corresponding to a certain acquisition time j, where j=1, 2, 3...n2, and n2 represents the total number of acquisition times; respectively execute S41 on the scraper machine color distribution map corresponding to the acquisition time j, the scraper machine color distribution map corresponding to the acquisition time j-1, the scraper machine color distribution map corresponding to the acquisition time j, and the scraper machine color distribution map corresponding to the acquisition time j+1 to obtain the front difference parts and the number of the front difference parts, the rear difference parts, and the number of the rear difference parts; and so on, the scraper machine color distribution map at each moment corresponds to the number of the front difference parts and the number of the rear difference parts. It should be noted that the scraper machine color distribution map corresponding to the start acquisition time has only the number of the rear difference parts, and the scraper machine color distribution map corresponding to the end acquisition time has only the number of the front difference parts, and the numbers of the front difference parts and the rear difference parts are recorded as r1 and r2 respectively;

[0062] S43: arbitrarily take the RGB value of a certain part in the scraper machine color distribution map corresponding to a certain acquisition moment, and perform difference calculation on the RGB value of the adjacent part to obtain the adjacent difference. When the adjacent difference is greater than the set adjacent part difference, the part is recorded as an adjacent difference. Similarly, the number of adjacent differences of each part in the scraper machine color distribution map is obtained. The number of adjacent differences of each part is compared and analyzed with the set number interval. When the number of adjacent differences is greater than the maximum value in the set number interval, the part is recorded as a high-difference part. When the number of adjacent differences is within the set number interval, the part is recorded as a medium-difference part. When the number of adjacent differences is less than the minimum value in the set number interval, the part is recorded as a slight-difference part. Count the number of high-difference parts, medium-difference parts and slight-difference parts in the scraper machine color distribution map, and record them as r3, r4 and r5 respectively.

[0063] S44: r1, r2, r3, r4 and r5 are set by the formula Calculate the color difference value RZ, where b1, b2, b3, b4, and b5 are the set proportional coefficients; send the generated color difference value to the image recognition output module;

[0064] S5: The texture feature analysis module extracts a scraper image and a standard scraper image from the image database, and uses local binary patterns to extract texture features of the scraper image and the standard scraper image to obtain texture feature information and standard texture feature information, wherein the texture feature information includes texture energy, texture contrast, and texture entropy; the standard texture feature information includes standard texture energy, standard texture contrast, and standard texture entropy;

[0065] The texture energy, texture contrast and texture entropy are respectively calculated by difference with the standard texture energy, standard texture contrast and standard texture entropy to obtain energy difference, contrast difference and entropy difference, and are recorded as l1, l2 and l3 respectively; l1, l2 and l3 are substituted into the set formula LZ = c1×l1+c2×l2+c3×l3 to calculate the texture contrast value LZ, where c1, c2 and c3 are the set proportional coefficients respectively; the generated texture contrast value is sent to the image recognition output module; it can be seen from the formula that the greater the energy difference, the greater the contrast difference, and the greater the entropy difference, the greater the texture contrast value, indicating that the texture gap between the scraper image texture and the standard scraper image is greater, and the potential risk of scraper failure and abnormality is greater;

[0066] S6: using an edge detection algorithm to extract the contours of the scraper machine image and the standard scraper machine image to obtain a scraper machine contour and a standard scraper machine contour, and respectively calculating the contour area and the standard contour area;

[0067] The scraper profile is overlapped and compared with the standard scraper profile to obtain the overlap area. When the overlap area = profile area = standard profile area, it means that the size of the scraper is normal and no operation is performed. When the overlap area ≠ profile area ≠ standard profile area, the difference between the standard profile area and the overlap area is calculated to obtain the profile difference value, which is recorded as LZ. The generated profile difference value is sent to the image recognition output module.

[0068] S7: The image recognition output module retrieves the risk level of the scraper and performs a comprehensive analysis with the color difference value, texture and contour difference value to determine the working status of the scraper, as follows:

[0069] S71: setting a risk coefficient corresponding to each risk level, matching the risk level of the scraper with all the set risk levels to obtain a corresponding risk coefficient, and recording it as FZ;

[0070] S72: The risk factor FZ, the color difference value RZ, the texture comparison value WZ, and the contour difference value LZ are calculated using the set formula FRZ = d1 × FZ + d2 × RZ + d3 × WZ + d4 × LZ to obtain the state value FRZ, where d1, d2, d3, and d4 are set proportional coefficients respectively;

[0071] S73: Compare and analyze the state value with the set state interval. When the state value is greater than the maximum value in the set state interval, it means that the scraper is operating at a faulty state, and a fault warning is generated and the scraper is recorded as a faulty scraper; when the state value is within the set state interval, it means that the scraper is operating abnormally, and an abnormality warning is generated and the scraper is recorded as an abnormal scraper; when the state value is less than the minimum value in the set state interval, it means that the scraper is operating normally, and no operation is performed; the fault warning signal and the faulty scraper, the abnormality warning signal and the abnormal scraper are recorded as image recognition results of the scraper, and the image recognition results of the scraper are sent to the intelligent management module; the scraper image is analyzed for color, texture and contour features by the color feature analysis module, the texture feature analysis module and the contour analysis module respectively to output the scraper image recognition result, which can more comprehensively analyze the scraper image and improve the accuracy of the scraper image recognition result;

[0072] S8: The intelligent management module implements corresponding execution strategies based on the image recognition results of the scraper, as follows:

[0073] When the scraper machine image recognition result is a fault warning, the number and location of the faulty scraper machine are retrieved, and the corresponding responsible employee is retrieved based on the number of the faulty scraper machine. A location acquisition instruction is sent to the responsible employee's mobile terminal to obtain the responsible employee's location. It should be noted that the mobile terminal includes but is not limited to smartphones, smart watches and tablets;

[0074] Retrieve the numbers and locations of the standby scrapers, calculate the distance between the standby scraper and the responsible employee, and the distance between the standby scraper and the faulty scraper, and sum the two to get the interval distance. Compare the interval distances of all standby scrapers, and record the standby scraper with the smallest interval distance as the target scraper. Generate several paths based on the locations of the target scraper and the faulty scraper, and select the shortest path as the target path.

[0075] The number and location of the faulty scraper, the number and location of the target scraper, and the target path are sent to the responsible employee's mobile terminal. After the responsible employee completes the replacement between the faulty scraper and the target scraper, the replacement completion signal is sent to the intelligent management module via the mobile terminal. When the intelligent management module receives the maintenance completion signal, it controls the standby scraper to start working.

[0076] When the scraper image recognition result is an abnormality warning, the number and location of the abnormal scraper will be retrieved, and the corresponding responsible employee will be retrieved based on the number of the abnormal scraper. A location acquisition instruction will be sent to the responsible employee's mobile terminal to obtain the responsible employee's location; the distance difference between the responsible employee's location and the scraper location will be calculated to obtain the distance;

[0077] Retrieve the responsible employee's historical maintenance times, as well as the historical distance and duration of each maintenance. It should be noted that the historical duration refers to the time it takes the responsible employee to reach the scraper's location after traveling the historical distance. Divide the historical distance by the historical duration to obtain the historical speed, and average all historical speeds to obtain the historical average speed.

[0078] Divide the distance between the routes by the historical average speed to obtain the estimated duration, and then control the scraper to continue working for a time equal to the estimated duration before stopping to cooperate with the responsible employee to perform scraper maintenance work; when the responsible employee completes the maintenance, a maintenance completion signal is sent to the intelligent management module through the mobile terminal. When the intelligent management module receives the maintenance completion signal, it controls the scraper to start working; the scraper is stopped by controlling the distance between the responsible employee and the scraper, which can avoid the responsible employee stopping the scraper in advance before reaching the abnormal warning position, resulting in unreasonable downtime, and further causing waste of production resources of the coal preparation plant; and achieve abnormal scraper maintenance while maximizing the working efficiency of the coal preparation plant.

[0079] Furthermore, the present invention processes the scraper maintenance moment and the corresponding maintenance duration to obtain a scraper maintenance report, calculates the time difference of adjacent maintenance moments to obtain the maintenance interval duration, compares and analyzes them to obtain long intervals, medium intervals and short intervals, and counts the number of long intervals, medium intervals and short intervals respectively; calculates the slope of the line segment connected by the midpoints of two adjacent equal-width strips, compares and analyzes the slope with zero to obtain the increasing slope and the decreasing slope, analyzes the relationship between adjacent slopes to obtain adjacent increasing slopes and adjacent decreasing slopes, and counts the number of groups of adjacent increasing slopes and the number of groups of adjacent decreasing slopes respectively; combines the use time and working time of the scraper with the number of long intervals, the number of medium intervals, the number of short intervals, the number of groups of adjacent increasing slopes and the number of groups of adjacent decreasing slopes for numerical analysis and processing to obtain the maintenance usage value, and accordingly obtains the monitoring interval duration and risk level; and obtains the monitoring requirements of the scraper by conducting in-depth analysis of the scraper information, so that the scraper can be monitored more reasonably and scientifically;

[0080] Furthermore, the present invention extracts color features from the scraper image to obtain a color distribution map of the scraper, and performs color feature analysis on the color distribution map of the scraper to obtain a color difference value; uses a local binary pattern to extract texture features from the scraper image and the standard scraper image to obtain texture feature information and standard texture feature information, and compares and analyzes the texture feature information and the standard texture feature information to obtain a texture comparison value; uses an edge detection algorithm to perform contour feature analysis on the scraper image and the standard scraper image to obtain a contour difference value; then retrieves the risk level of the scraper, and performs a comprehensive analysis on it with the color difference value, texture and contour difference value to determine the working state of the scraper, and outputs the image recognition result of the scraper; performs color, texture and contour feature analysis on the scraper image respectively through the color feature analysis module, the texture feature analysis module and the contour analysis module to output the scraper image recognition result, which can more comprehensively analyze the scraper image and improve the accuracy of the scraper image recognition result;

[0081] Furthermore, the present invention implements corresponding execution strategies through the image recognition results of the scraper. When the image recognition results are a fault warning and a faulty scraper, the faulty scraper number and position are retrieved, and a spare scraper and path are selected, and sent to the responsible employee to replace the faulty scraper with the spare scraper to ensure the safe operation of the scraper; when the image recognition results are an abnormal warning and an abnormal scraper, the abnormal scraper number and position, as well as the position of the responsible employee are retrieved, and analyzed based on the historical maintenance situation of the responsible employee to obtain the estimated duration, and the scraper is controlled to continue working for a time equal to the estimated duration before shutting down to cooperate with the responsible employee to perform scraper maintenance work; the scraper is controlled to stop by the distance between the responsible employee and the scraper, which can avoid the responsible employee from stopping in advance before reaching the location of the abnormal warning scraper, resulting in unreasonable downtime, and further causing waste of production resources of the coal preparation plant; and the abnormal scraper is repaired while maximizing the work efficiency of the coal preparation plant.

[0082] Thus far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art may make equivalent changes or substitutions to the relevant technical features, and the technical solutions after such changes or substitutions will fall within the scope of protection of the present invention.

[0083] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that the present invention is susceptible to various modifications and variations. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be within the scope of protection of the present invention.

Claims

1. A scraper image recognition method based on image processing, characterized in that: The following steps are involved: S1: Collect scraper information and scraper image information; S2: Analyze scraper information to obtain monitoring intervals and risk levels, where the risk levels include high-risk scraper, medium-risk scraper, and low-risk scraper; S3: Monitor the scraper according to the monitoring interval; S4: Extract the scraper machine image and perform color feature extraction on the scraper machine image to obtain a scraper machine color distribution map. Perform color feature analysis based on the scraper machine color distribution map to obtain a color difference value. The specific steps are as follows: S41: Identify the scraper color distribution map to obtain the RGB value corresponding to each part in the image, and compare and analyze the RGB values of the same part in the scraper images at adjacent acquisition moments to obtain the difference parts; S42: randomly select a scraper machine color distribution map corresponding to a certain sampling time, and perform S41 on it together with the scraper machine color distribution maps corresponding to the two adjacent sampling times to obtain the front difference part and the number of the front difference parts, the rear difference part and the number of the rear difference parts; S43: randomly taking the RGB value of a certain part in the scraper machine color distribution map corresponding to a certain acquisition moment, and performing difference calculation on the RGB value of the part and the adjacent part to obtain an adjacent difference. When the adjacent difference is greater than the set adjacent part difference, the part is recorded as an adjacent difference. Similarly, the number of adjacent differences of each part in the scraper machine color distribution map is obtained; the number of adjacent differences of each part is compared and analyzed with the set number interval to obtain high-difference parts, medium-difference parts and slight-difference parts; and the number of high-difference parts, medium-difference parts and slight-difference parts in the scraper machine color distribution map is counted. S44: performing a formula calculation and analysis on the number of front difference parts, the number of rear difference parts, the number of high difference parts, the number of medium difference parts, and the number of slight difference parts to obtain a color difference value; S5: extracting the scraper image and performing texture feature extraction on the image to obtain texture feature information, and performing texture analysis on the texture feature information to obtain a texture difference value; S6: extracting the scraper machine image and the standard scraper machine image and performing contour feature extraction on them to obtain the scraper machine contour and the standard scraper machine contour, and performing contour analysis on the two to obtain a contour difference value; S7: Comprehensively analyzing the risk level, color difference value, texture difference value, and contour difference value of the scraper to output an image recognition result of the scraper; S8: Pre-execute a corresponding strategy for the scraper according to the scraper image recognition result.

2. The scraper machine image recognition method based on image processing according to claim 1, characterized in that: The scraper is analyzed to obtain the monitoring interval and risk level. The specific steps are as follows: 201: Get the working time and service time of the scraper; 202: Obtain the number of maintenance times of the scraper, the maintenance time and the maintenance duration of each maintenance, use the maintenance time as the horizontal axis, draw vertical bars of equal width at the corresponding maintenance time, and the length of the vertical bars of equal width is equal to the corresponding maintenance duration in value to create a scraper maintenance report; 203: Calculate the difference between the maintenance times corresponding to two adjacent maintenance times to obtain the maintenance interval duration, compare the interval maintenance duration with the set duration interval to obtain long intervals, medium intervals, and short intervals, and count the number of long intervals, medium intervals, and short intervals respectively; 204: Calculate the slope of the line segment connecting the midpoints of two adjacent equal-width strips. When the slope is greater than zero, record it as an increasing slope; when the slope is less than zero, record it as a decreasing slope. When any two increasing slopes have overlapping maintenance moments, the two increasing slopes are recorded as a group of adjacent increasing slopes, and the number of groups of adjacent increasing slopes is counted. When any two decreasing slopes have overlapping maintenance moments, the two decreasing slopes are recorded as a group of adjacent decreasing slopes, and the number of groups of adjacent decreasing slopes is counted. 205: The working hours, usage hours, number of long intervals, number of medium intervals, number of short intervals, number of groups of adjacent increasing slopes, and number of groups of adjacent decreasing slopes are numerically analyzed to obtain a maintenance usage index; the maintenance usage index is compared with the set index range to obtain high-risk scraper, medium-risk scraper, and low-risk scraper.

3. The scraper image recognition method based on image processing according to claim 1, characterized in that: Texture analysis is performed on the texture feature information to obtain the texture difference value. The specific steps are as follows: Extracting a scraper machine image and a standard scraper machine image, and performing texture feature extraction on the scraper machine image and the standard scraper machine image using a local binary pattern to obtain texture feature information and standard texture feature information, wherein the texture feature information includes texture energy, texture contrast, and texture entropy; and the standard texture feature information includes standard texture energy, standard texture contrast, and standard texture entropy; The texture energy, texture contrast and texture entropy are respectively calculated to obtain the energy difference, contrast difference and entropy difference with the standard texture energy, standard texture contrast and standard texture entropy, and the three are processed through numerical analysis to obtain the texture contrast value.

4. The scraper machine image recognition method based on image processing according to claim 1, characterized in that: Perform contour analysis on the scraper machine image and the standard scraper machine image to obtain the contour difference value. The specific steps are as follows: The scraper machine image and the standard scraper machine image are subjected to contour extraction using an edge detection algorithm to obtain the scraper machine contour and the standard scraper machine contour, and the contour area and the standard contour area are calculated; The scraper profile is overlapped and compared with the standard scraper profile to obtain the overlap area. When the overlap area ≠ profile area ≠ standard profile area, the difference between the standard profile area and the overlap area is calculated to obtain the profile difference value.

5. The scraper machine image recognition method based on image processing according to claim 1, characterized in that: The specific steps for outputting the image recognition results of the scraper are: Set different risk levels to correspond to a risk coefficient, and match the risk level of the scraper with all the set risk levels to obtain the corresponding risk coefficient; The risk coefficient, color difference value, texture comparison value and contour difference value are normalized and their values are taken, and the values are analyzed to obtain the status value; Compare and analyze the state value with the set state interval to generate fault warning and abnormal warning, and record the scraper machines corresponding to the fault warning and abnormal warning as faulty scraper machine and abnormal scraper machine respectively; The fault warning signal and the faulty scraper, the abnormal warning signal and the abnormal scraper are recorded as the image recognition results of the scraper.

6. The scraper machine image recognition method based on image processing according to claim 1, characterized in that: According to the scraper machine image recognition results, the scraper machine performs the corresponding pre-execution strategy. The specific steps are as follows: When the scraper machine image recognition result is a fault warning, the number and location of the faulty scraper machine are retrieved, and the corresponding responsible employee and the location of the responsible employee are retrieved based on the number of the faulty scraper machine; the target scraper machine and the target path are selected from the spare scrapers, and the number and location of the faulty scraper machine, the number and location of the target scraper machine, and the target path are sent to the responsible employee's mobile terminal; When the scraper machine image recognition result is an abnormality warning, the number and location of the abnormal scraper machine will be retrieved, and the corresponding responsible employee and the location of the responsible employee will be retrieved based on the number of the abnormal scraper machine; the distance difference between the location of the responsible employee and the location of the scraper machine will be calculated to obtain the distance; Retrieve the responsible employee's historical maintenance times, as well as the historical distance and duration of each maintenance, and analyze them to obtain the historical average speed; Divide the distance between the routes by the historical average speed to obtain the estimated duration. Then control the scraper to continue working for a time equal to the estimated duration before stopping.

Images