Mining intrinsic safety type radar coal piling sensor with echo processing module

Abstract

The invention relates to the technical field of coal piling quantity measurement, in particular to a mining intrinsic safety type radar coal piling sensor with an echo processing module, which comprises a radar monitoring layer, a machine vision monitoring layer, a data processing and control layer, a network communication layer, a display layer, a sound-light alarm layer and a power supply layer. According to the invention, the radar monitoring layer effectively overcomes the interference of dust, water vapor and the like through the echo processing unit and accurately detects the coal piling condition, the machine vision monitoring layer processes the coal piling area image to make the image clearer, and the data processing and control layer performs normalization processing and data synchronization processing on the two kinds of data to make the data more clear. The coal piling condition and the coal piling height are comprehensively judged in combination with the preset threshold value, the coal piling condition can be comprehensively and accurately monitored and analyzed in real time, accurate monitoring and timely response to the coal piling condition are achieved, and reliable safety guarantee is provided for coal mine production.

Description

Technical Field

[0001] The present invention relates to the technical field of measuring the coal stacking amount, and particularly relates to a mine intrinsically safe radar coal stacking sensor with an echo processing module. Background Art

[0002] In the process of coal mine production, the belt conveyor is a key equipment for coal transportation. However, due to factors such as large coal production, fast transportation speed, and complex working environment, the belt conveyor is prone to coal stacking. Coal stacking not only affects the normal transportation of coal, but may also cause belt deviation, tearing, and even lead to serious safety accidents such as motor overload and fire, bringing huge economic losses and safety hazards to coal mine production. At present, most coal stacking sensors are used to monitor the coal stacking situation on the belt conveyor to protect the conveyor.

[0003] At present, the common coal stacking sensors on the market mainly include ultrasonic coal stacking sensors and infrared coal stacking sensors. In an environment with a large amount of coal dust, the ultrasonic signals of the ultrasonic coal stacking sensors are easily interfered, resulting in a decrease in detection accuracy. When there is moisture or other impurities on the surface of the coal pile, it affects the reflection effect of the infrared light of the infrared coal stacking sensors, thereby affecting the accuracy of detection. Moreover, most of the existing coal stacking sensors have a single function, cannot comprehensively and accurately monitor and analyze the coal stacking situation in real time, and cannot achieve real-time intelligent control, which poses a potential threat to the safety of coal mine production. Summary of the Invention

[0004] Aiming at the problems existing in the prior art, the purpose of the present invention is to provide a mine intrinsically safe radar coal stacking sensor with an echo processing module.

[0005] To achieve the above purpose, the present invention provides the following technical solutions: A mine intrinsically safe radar coal stacking sensor with an echo processing module, comprising a radar monitoring layer, a machine vision monitoring layer, a data processing and control layer, a network communication layer, a display layer, an audible and visual alarm layer, and a power supply layer: The radar monitoring layer uses millimeter-wave radar technology to transmit radar signals to the measured coal stacking area and receive the radar echo signals reflected from the surface of the coal stack, performs filtering processing, mixing processing, and digital signal processing on the radar echo signals, extracts the frequency characteristic information and phase characteristic information of the echo signals, and obtains the distance data of the measured coal stacking area; The machine vision monitoring layer uses an industrial camera to capture images of the coal pile area, and performs denoising processing, defogging and enhancement processing, grayscale processing, edge detection, three-frame difference method processing, and binarization processing on the images to obtain the machine vision processing results; The data processing and control layer performs normalization processing on the distance data of the measured coal stockpiling area and the data of the machine vision processing results, combines the preset threshold values, conducts data fusion and comprehensive judgment to determine whether there is a coal stockpiling situation and the height of the coal stockpile. When it is judged that the coal stockpiling is abnormal, an abnormal signal is transmitted to the display layer and the coal mine terminal server through the network communication layer, and the coal mine terminal server automatically controls the conveyor to stop.

[0006] Preferably, the network communication layer is used to receive data from each layer, for data transmission and aggregation, and transmits the data to the coal mine terminal server through a 4G / 5G or ZigBee wireless communication module. The display layer receives the coal stockpile data of the data processing and control layer and displays it through an LED display screen. The acoustic-optic alarm layer receives the abnormal signal of the data processing and control layer and issues an acoustic-optic alarm prompt. The power supply layer uses a converter to convert the intrinsically safe power input from the outside into the stable DC voltage required by each layer to supply power to each layer of the intrinsically safe mine radar coal stockpile sensor.

[0007] Preferably, the radar monitoring layer includes a radar transmitting module, a radar receiving module, a preliminary amplification module, an echo processing unit, and a radar data processing module; The radar transmitting module uses millimeter-wave radar technology to transmit radar signals to the measured coal stockpiling area; The radar receiving module receives the radar echo signals reflected from the surface of the coal stockpile; The preliminary amplification module preliminarily amplifies the radar echo signals through a low-noise amplifier; The echo processing unit extracts the frequency characteristic information and phase characteristic information of the echo signals through filtering processing, mixing processing, and digital signal processing to obtain the distance data of the measured coal stockpiling area; The radar data processing module uses a moving average filtering algorithm to perform average processing on multiple distance data of the measured coal stockpiling area continuously collected by the echo processing unit to obtain the coal stockpile height value.

[0008] Preferably, the echo processing unit includes a signal filtering module, a mixing module, and a digital signal processing module; The signal filtering module uses a band-pass filter to filter out the high-frequency noise and low-frequency interference signals in the echo signals and retains the frequency components related to the target reflection signals; The mixing module mixes the filtered echo signals with the local oscillation signals to convert the high-frequency echo signals into intermediate-frequency signals; The digital signal processing module performs spectrum analysis on the intermediate-frequency signals, extracts the frequency characteristic information and phase characteristic information of the echo signals, and obtains the distance data of the measured coal stockpiling area.

[0009] Preferably, the digital signal processing module performs spectral analysis on the intermediate frequency signal to extract the frequency characteristic information and phase characteristic information of the echo signal, including the following steps: Sampling and quantization: Sampling and quantizing the intermediate frequency signal to convert the analog intermediate frequency signal into a digital signal. After sampling, the signal is quantized to convert the continuous analog amplitude into discrete digital values; Fast Fourier transform: Performing a fast Fourier transform on the sampled and quantized digital signal. Through fast Fourier transform analysis, the spectral characteristics of the echo signal, that is, the distribution of the signal at different frequencies, are obtained. By analyzing the spectrum, the frequency peaks related to the coal pile reflection signal are found; Feature extraction and analysis: By analyzing the phase and amplitude characteristics of the echo signal, the frequency characteristic information and phase characteristic information of the echo signal are extracted to obtain the distance data of the measured coal pile area.

[0010] Preferably, the machine vision monitoring layer includes an image acquisition module, an image denoising module, an image defogging and enhancement processing module, a grayscale processing module, an edge detection module, a three-frame difference method processing module, and a binarization processing module; The image acquisition module uses an industrial camera to capture images of the coal pile area; The image denoising module receives the image data and reads the video frames, processes the image using the wavelet denoising method to obtain wavelet coefficients in different directions and scales, performs threshold processing through the threshold denoising method based on wavelet transform (VisuShrink), estimates a reasonable threshold, sets the wavelet coefficients smaller than the threshold to 0, and retains the others. According to the calculation result of the threshold estimation, the processed wavelet coefficients are further inversely transformed to obtain the denoised coal pile image. The functional expression of the VisuShrink threshold is as follows: ; In the formula: A represents the threshold, represents the standard deviation of the image noise, B represents the product of the image width and height, represents the median function, represents the wavelet function at the minimum layer; The image defogging and enhancement processing module uses the dark channel image defogging and enhancement algorithm to perform defogging and enhancement processing on the denoised coal pile image to obtain the defogged image; The grayscale processing module grayscales the defogged image using the weighted average method. The image grayscale value is expressed as: ; In the formula: represents the red component in the image, represents the green component in the image, Represents the blue component in the image; The edge detection module uses the Canny operator to detect edges in the grayscale image, and calculates the direction of each pixel point in the image and gradient , performs non-maximum suppression on the gradient magnitude to obtain an edge intensity map, accurately locates the edges, and then uses the double-threshold algorithm to refine and connect the edges to obtain an edge image: ; In the formula: represents the gradient of the pixel point in the horizontal direction, represents the gradient of the pixel point in the vertical direction; The three-frame difference method processing module uses the difference algorithm to analyze adjacent frame images, compares the grayscale values of pixel points at the same position, calculates their differences, combines adjacent pixel points for connectivity analysis, and if the absolute value of the difference exceeds the set threshold, determines these changes as moving objects and forms a complete moving object contour to achieve the extraction of moving objects; The binarization processing module further converts the image into an image with only black and white values to obtain the machine vision processing result.

[0011] Preferably: The data processing and control layer includes a normalization module, a data fusion and judgment module, and a self-check module; The normalization module receives the distance data of the measured coal pile area output by the radar monitoring layer and the machine vision processing result data output by the machine vision monitoring layer, and performs normalization processing on them; The data fusion and judgment module fuses and judges the two sets of data after normalization processing to judge whether there is a coal pile situation and the coal pile height; The self-check module performs system self-check after the sensor is powered on, checks whether the hardware connections of each module are normal and whether the software program can run normally. If an abnormality is found during the self-check process, the data is uploaded to the display layer through the network communication layer for display.

[0012] Preferably: The data fusion and judgment module fuses and judges the two kinds of data after normalization processing, including the following steps: B1. Synchronize the original data of the distance data characteristics of the measured coal pile area and the machine vision processing result data characteristics, that is, spatial synchronization and time synchronization; B2. Combine the preset threshold, independently judge the two sets of synchronized data, and use the support vector machine classifier trained based on distance characteristics to judge whether the coal pile distance reaches the threshold to obtain a decision result , and use the neural network classifier trained based on image characteristics Judge whether the distance of the coal pile reaches the threshold to obtain the decision result ; B3. Assign different weights according to the importance of the distance feature and the image feature when judging the distance, that is and , and use the weighted voting method to fuse the two decision results and : ; In the formula: and are binary classification results. When the distance of the coal pile reaches the threshold, it is 1, indicating that coal piling and the height of the coal pile occur. When it does not reach the threshold, it is 0, indicating that no coal piling occurs; B4. Model training and evaluation: Use the fused feature data to train the classification model. The classification model can be a logistic regression model, a random forest model or a deep learning model. The K-fold cross-validation method is used to evaluate the trained model. The data set is divided into K parts, and each time one part is used as the test set, and the remaining K-1 parts are used as the training set. Repeat K times, calculate the evaluation indicators each time, that is, accuracy, recall rate and F1 value, and take the average value as the final evaluation result; B5. The data fusion and judgment module uploads the fusion and judgment data to the display layer and the coal mine terminal server through the network communication layer; B6. When the data fusion and judgment module detects abnormal coal piling, it transmits the abnormal signal to the display layer and the coal mine terminal server through the network communication layer, and the coal mine terminal server automatically controls the conveyor to stop.

[0013] Beneficial effects Adopting the technical solution provided by the present invention, compared with the known public technology, it has the following beneficial effects: (1) In the present invention, the radar monitoring layer uses millimeter-wave radar technology and an advanced echo processing unit to obtain the dynamic change data of the coal pile height in real time. The signal filtering, mixing and digital signal processing technologies in the echo processing unit can accurately distinguish useful echo signals and interference signals, effectively overcome interference such as dust and water vapor, accurately detect coal piling, have strong adaptability to harsh environments such as coal dust, greatly improve the anti-interference ability of the sensor in a complex environment, reduce the occurrence of false alarms, and improve the accuracy of detection; (2) In the present invention, the machine vision monitoring layer uses an industrial camera to capture images of the coal pile area. The image denoising module uses wavelet denoising method to denoise the images, effectively removing the noise in the images while retaining the detailed information of the images. The image dehazing and enhancement processing module uses the dark channel image dehazing and enhancement algorithm to dehaze and enhance the images, which can make the images clearer, basically removing the influence of dust on the images, effectively improving the image quality. The edge detection module uses the Canny operator to detect the edges of the grayscale images, which can effectively suppress the noise in the images, reduce the data volume, highlight the important image features, simplify the subsequent image processing tasks, strengthen the contour of the image edges, and obtain clearer machine vision processing results, facilitating the subsequent judgment of the coal pile situation. (3) In the present invention, the data processing and control layer performs normalization processing and data synchronization processing on the measured distance data of the coal pile area and the machine vision processing result data, and then combines the preset thresholds to independently judge the two synchronized data to obtain decision results. The weighted voting method is used to fuse the two decision results, so as to comprehensively judge whether there is a coal pile situation and the height of the coal pile, which can comprehensively and accurately monitor and analyze the coal pile situation in real time, and can perform real-time intelligent control on the coal pile situation, realizing accurate monitoring and timely response to the coal pile situation, providing a reliable safety guarantee for coal mine production. Description of the Drawings

[0014] Figure 1 It is a system block diagram of a mine intrinsically safe radar coal pile sensor with an echo processing module.

[0015] In the figure: 1. Radar monitoring layer; 2. Machine vision monitoring layer; 3. Data processing and control layer; 4. Network communication layer; 5. Display layer; 6. Acousto-optic alarm layer; 7. Power supply layer. Detailed Embodiment

[0016] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0017] The following further describes the present invention with reference to embodiments.

[0018] Please refer to Figure 1 As shown, a mine intrinsically safe radar coal pile sensor with an echo processing module includes a radar monitoring layer 1, a machine vision monitoring layer 2, a data processing and control layer 3, a network communication layer 4, a display layer 5, an acousto-optic alarm layer 6, and a power supply layer 7.

[0019] The radar monitoring layer 1 uses millimeter-wave radar technology to transmit radar signals to the measured coal stockpiling area and receive the radar echo signals reflected from the surface of the coal stockpile. The radar echo signals are subjected to filtering, mixing, and digital signal processing to extract the frequency characteristic information and phase characteristic information of the echo signals, and the distance data of the measured coal stockpiling area is obtained. Specifically, the radar monitoring layer 1 includes a radar transmitting module, a radar receiving module, a preliminary amplification module, an echo processing unit, and a radar data processing module.

[0020] Wherein: The radar transmitting module uses millimeter-wave radar technology to transmit radar signals to the measured coal stockpiling area. The radar transmitting module can transmit high-frequency and narrow-beam radar signals to the measured area. By installing a magnetic ring antenna with good anti-interference ability and high-precision ranging performance obliquely above the coal stockpiling position on the belt conveyor, it is ensured that the omnidirectional state parameters of the coal pile can still be accurately obtained in a complex environment, so that the transmitted signal has good directivity, can concentrate energy on the target area on the conveyor, and reduce signal scattering and interference.

[0021] The radar receiving module receives the radar echo signals reflected from the surface of the coal stockpile. By using a high-sensitivity receiving antenna and a low-noise amplifier, weak echo signals can be effectively captured. In this embodiment, it is preferably that the receiving antenna and the transmitting antenna adopt an isolation design to reduce the interference of the transmitted signal on the received signal.

[0022] The preliminary amplification module preliminarily amplifies the radar echo signals through a low-noise amplifier.

[0023] The echo processing unit includes a signal filtering module, a mixing module, and a digital signal processing module.

[0024] The signal filtering module uses a band-pass filter. The band-pass filter has a specific passband. Only signals with frequencies within this passband range can pass through, while signals higher or lower than this passband are greatly attenuated. By using the band-pass filter, high-frequency noise and low-frequency interference signals in the echo signals can be effectively filtered out, and the frequency components related to the target reflection signals are retained, improving the accuracy of subsequent processing.

[0025] The mixing module mixes the filtered echo signals with the local oscillation signals to convert the high-frequency echo signals into intermediate-frequency signals. The intermediate-frequency signals have lower frequencies, which are convenient for subsequent signal processing and analysis. At the same time, it also reduces the performance requirements for subsequent processing circuits and is convenient for subsequent processing, including the following steps: The local oscillation signal generator generates a stable signal with a frequency of ; When the filtered echo signal and the local oscillation signal When they are simultaneously input into the mixer, the mixer multiplies these two signals by using its non-linear characteristics; According to the product formula of trigonometric functions , the output signal after mixing contains two new frequency components: the sum frequency and the difference frequency , where is the frequency of the echo signal; Through subsequent filters, the difference frequency signal is selected as the intermediate frequency signal for output.

[0026] The digital signal processing module performs spectrum analysis on the intermediate frequency signal, extracts the frequency characteristic information and phase characteristic information of the echo signal, accurately distinguishes the useful echo signal from the interference signal, and obtains the distance data of the measured coal pile area, including the following steps: Sampling and quantization: Sampling and quantizing the intermediate frequency signal, converting the analog intermediate frequency signal into a digital signal. After sampling, the signal undergoes quantization processing to convert the continuous analog amplitude into discrete digital values; Fast Fourier transform: Performing a fast Fourier transform on the sampled and quantized digital signal. Through fast Fourier transform analysis (FFT analysis), the spectral characteristics of the echo signal are obtained, that is, the distribution of the signal at different frequencies. By analyzing the spectrum, the frequency peak related to the coal pile reflection signal is found; Feature extraction and analysis: By analyzing the phase and amplitude characteristics of the echo signal, the frequency characteristic information and phase characteristic information of the echo signal are extracted to obtain the distance data of the measured coal pile area.

[0027] The radar data processing module adopts a moving average filtering algorithm to average multiple distance data of the measured coal pile area continuously collected by the echo processing unit to obtain the coal pile height value.

[0028] Radar monitoring layer 1, by adopting millimeter-wave radar technology and an advanced echo processing unit, calculates the time difference between the transmitted signal and the echo signal, and combines the propagation speed of millimeter waves in the air to accurately measure the distance between the coal pile and the radar. Radar monitoring layer 1 can obtain the dynamic change data of the coal pile height in real time. The signal filtering, mixing, and digital signal processing technologies in the echo processing unit can accurately distinguish the useful echo signal from the interference signal, effectively overcome the interference of dust, water vapor, etc., accurately detect the coal pile situation, have strong adaptability to harsh environments such as coal dust, greatly improve the anti-interference ability of the sensor in complex environments, reduce the occurrence of false alarms, and improve the detection accuracy.

[0029] The machine vision monitoring layer 2 uses an industrial camera to capture images of the coal pile area, and performs denoising processing, defogging and enhancement processing, grayscale processing, edge detection, three-frame difference method processing, and binarization processing on the images to obtain the machine vision processing results. Specifically, the machine vision monitoring layer 2 includes an image acquisition module, an image denoising module, an image defogging and enhancement processing module, a grayscale processing module, an edge detection module, a three-frame difference method processing module, and a binarization processing module.

[0030] Among them: The image acquisition module uses an industrial camera to capture images of the coal pile area. A dust-proof and explosion-proof industrial camera is selected, equipped with a high-resolution image sensor, which can clearly image in a dim environment. The camera can cover the area where coal may be piled on the belt conveyor.

[0031] The image denoising module receives image data and reads video frames, processes the image using the wavelet denoising method to obtain wavelet coefficients in different directions and scales, performs threshold processing through the threshold denoising method (VisuShrink) based on wavelet transform, estimates a reasonable threshold, sets the wavelet coefficients smaller than the threshold to 0, and retains the others. According to the calculation result of the threshold estimation, the processed wavelet coefficients are further inverse-transformed to obtain the denoised coal pile image. The function expression of the VisuShrink threshold is as follows: ; In the formula: A represents the threshold, represents the standard deviation of the image noise, B represents the product of the width and height of the image, represents the median function, represents the wavelet function at the minimum layer; Among them, wavelet denoising is an image denoising method based on wavelet analysis. The image signal is decomposed into wavelet coefficients at different scales, and then the noise is removed through threshold processing to finally obtain the denoised coal pile image. Through wavelet denoising, the noise in the image can be effectively removed while retaining the detail information of the image, avoiding the image blurring that may be caused by traditional linear filters.

[0032] The image defogging and enhancement processing module uses the dark channel image defogging and enhancement algorithm to perform defogging and enhancement processing on the denoised coal pile image to obtain the defogged image, including the following steps: A1. Represent the denoised coal pile image as: ; In the formula: x is the image pixel point, is the foggy image, is the defogged image, F is the atmospheric light value of the current scene, is the scene transmittance, which is used to describe the part of light that is not scattered during the process of passing through the medium and reaching the imaging device; A2. The following formula can be obtained according to the above formula: ; A3. The basic principle of the dark channel image defogging and enhancement algorithm is that in a fog-free image, at least one of the three channels of a pixel point has a very low brightness, even close to 0, that is: ; In the formula: y is the position of the pixel point, is the dark channel map of the image; A4. Substituting the formula in step A3 into the formula in step A1, we can get: ; In the formula: the image transmittance is ; A5. By increasing the parameter , The larger the value, the smaller the transmittance, and the more thorough the image defogging. In this embodiment, the value of is preferably set to 0.95; Among them, using the dark channel image defogging and enhancement algorithm to perform defogging and enhancement processing on the denoised coal heap image can make the image clearer, basically remove the influence of dust on the image, and effectively improve the image quality.

[0033] The grayscale processing module grayscales the defogged image using the weighted average method. The image grayscale value is expressed as: ; In the formula: represents the red component in the image, represents the green component in the image, represents the blue component in the image.

[0034] The edge detection module uses the edge detection algorithm (Canny), that is, the Canny operator to perform edge detection on the grayscaled image, calculates the direction and gradient of each pixel point in the image, suppresses the non-maximum value of the gradient amplitude to obtain the edge intensity map, accurately locates the edge, and then uses the double-threshold algorithm to refine and connect the edge to obtain the edge image: ; In the formula: represents the gradient of the pixel point in the horizontal direction, represents the gradient of the pixel point in the vertical direction; Among them, the Canny operator is used to perform edge detection on the grayscale image. The Canny edge detection operator has good edge localization ability, can effectively suppress the noise in the image, strengthen the contour of the image edge, and make the image clearer.

[0035] The three-frame difference method processing module uses the difference algorithm to analyze adjacent frame images. By comparing the grayscale values of pixel points at the same position and calculating their differences, adjacent pixel points are combined for connectivity analysis. If the absolute value of the difference exceeds the set threshold, these changes are determined as moving targets and a complete moving target contour is formed to achieve the extraction of moving targets; the binarization processing module further converts the image into an image with only black and white values to obtain the machine vision processing result.

[0036] The data processing and control layer 3 performs normalization processing on the distance data of the measured coal heap area and the machine vision processing result data, combines the preset threshold, performs data fusion and comprehensive judgment to determine whether there is a coal heap situation and the coal heap height. When it is judged that the coal heap is abnormal, the abnormal signal is sent to the display layer 5 and the coal mine terminal server through the network communication layer 4, and the coal mine terminal server automatically controls the conveyor to stop. Specifically, the data processing and control layer 3 includes a normalization module, a data fusion and judgment module, and a self-check module.

[0037] The normalization module receives the distance data of the measured coal heap area output by the radar monitoring layer 1 and the machine vision processing result data output by the machine vision monitoring layer 2, and performs normalization processing on them to unify the numerical ranges of the distance data of the measured coal heap area and the machine vision processing result data.

[0038] The data fusion and judgment module fuses and judges the two kinds of data after normalization processing to determine whether there is a coal heap situation and the coal heap height, including the following steps: B1. Synchronize the original data of the distance data characteristics of the measured coal heap area and the machine vision processing result data characteristics, that is, spatial synchronization and time synchronization; B2. Combine the preset threshold to independently judge the two groups of synchronized data, and use the support vector machine classifier trained based on distance characteristics to judge whether the coal heap distance reaches the threshold to obtain the decision result , and use the neural network classifier trained based on image characteristics to judge whether the coal heap distance reaches the threshold to obtain the decision result ; B3. Assign different weights according to the importance of distance characteristics and image characteristics when judging distance, that is and , and use the weighted voting method for the two decision results and Fusion: ; In the formula: and are the binary classification results. When the coal pile distance reaches the threshold value of 1, it is judged that the coal piling situation and the coal piling height occur; when it does not reach 0, it is judged that the coal piling situation does not occur. B4. Model training and evaluation: Use the fused feature data to train the classification model. The classification model can be a logistic regression model, a random forest model, or a deep learning model. The K-fold cross-validation method is used to evaluate the trained model. The data set is divided into K parts. Each time, one part is used as the test set, and the remaining K - 1 parts are used as the training set. Repeat K times, calculate the evaluation metrics each time, namely accuracy, recall rate, and F1 value, and take the average as the final evaluation result. B5. The data fusion and judgment module uploads the fusion and judgment data to the display layer 5 and the coal mine terminal server through the network communication layer 4. B6. When the data fusion and judgment module detects coal piling anomalies, it transmits the anomaly signal to the display layer 5 and the coal mine terminal server through the network communication layer 4. The coal mine terminal server automatically controls the conveyor to stop.

[0039] The self-check module performs system self-check after the sensor is powered on, checks whether the hardware connections of each module are normal and whether the software program can run normally. If anomalies are found during the self-check process, the data is uploaded to the display layer 5 through the network communication layer 4 for display.

[0040] The network communication layer 4 is used to receive data from each layer, for data transmission and aggregation, and transmits the data to the coal mine terminal server through the 4G / 5G or ZigBee wireless communication module.

[0041] The display layer 5 receives the coal pile data of the data processing and control layer 3 and displays it through the LED display screen, which can intuitively display the detection results of the coal piling sensor, facilitating coal mine workers to discover and handle in a timely manner.

[0042] The audible and visual alarm layer 6 receives the anomaly signal from the data processing and control layer 3 and issues an audible and visual alarm prompt.

[0043] The power supply layer 7 uses a converter to convert the intrinsically safe power supply input externally into the stable DC voltage required by each layer, supplies power to each layer of the intrinsically safe mine radar coal piling sensor. At the same time, the power supply layer 7 also has overvoltage and overcurrent protection functions to prevent damage to the sensor caused by power supply failures and ensure the safe and reliable operation of the equipment underground.

[0044] In the present invention, during the operation of the belt conveyor, the radar monitoring layer 1 continuously emits millimeter-wave signals and receives reflected signals in real time. The echo processing unit processes the amplified radar echo signals through a signal filtering module, a mixing module, and a digital signal processing module. The signal filtering module can effectively filter out high-frequency noise and low-frequency interference signals in the echo signals through a band-pass filter, retain the frequency components related to the target reflection signals, and improve the accuracy of subsequent processing. The mixing module mixes the filtered echo signals with the local oscillation signals to convert the high-frequency echo signals into intermediate-frequency signals for subsequent processing. The digital signal processing module performs spectral analysis on the intermediate-frequency signals, extracts the frequency characteristic information and phase characteristic information of the echo signals, accurately distinguishes useful echo signals from interference signals, and obtains the distance data of the measured coal stacking area. By adopting millimeter-wave radar technology and an advanced echo processing unit, the radar monitoring layer 1 accurately measures the distance between the coal pile and the radar by calculating the time difference between the transmitted signal and the echo signal and combining the propagation speed of millimeter waves in the air. The radar monitoring layer 1 can obtain the dynamic change data of the coal pile height in real time. The signal filtering, mixing, and digital signal processing technologies in the echo processing unit can accurately distinguish useful echo signals from interference signals, effectively overcome interferences such as dust and water vapor, accurately detect the coal stacking situation, have strong adaptability to harsh environments such as coal dust, greatly improve the anti-interference ability of the sensor in complex environments, reduce the occurrence of false alarms, and improve the accuracy of detection. The machine vision monitoring layer 2 uses an industrial camera to capture images of the coal pile area. The image denoising module performs denoising processing on the images. By using the wavelet denoising method to process the images, wavelet coefficients in different directions and scales are obtained, and the denoised coal stacking images are obtained. Through wavelet denoising, the noise in the images can be effectively removed while retaining the detail information of the images, avoiding image blurring that may be caused by traditional linear filters. Then, through the image dehazing and enhancement processing module, the denoised coal stacking images are subjected to dehazing and enhancement processing using the dark channel image dehazing and enhancement algorithm, which can make the images clearer, basically remove the influence of dust on the images, effectively improve the image quality. The grayscale processing module grayscales the dehazed images using the weighted average method. The edge detection module uses the Canny operator to perform edge detection on the grayscaled images, which can effectively suppress the noise in the images, reduce the data volume, highlight important image features, simplify the subsequent image processing tasks, and strengthen the contour of the image edges, making the images clearer. The three-frame difference method processing module performs three-frame difference method processing on the images to extract moving objects. The binarization processing module further converts the images into images with only two values of black and white to obtain the machine vision processing results. The data processing and control layer 3 performs normalization processing and data synchronization processing on the distance data of the coal pile area to be measured and the data of the machine vision processing results, and then combines the preset threshold to independently judge the two synchronized data, and uses the support vector machine (SVM) classifier trained based on distance features to judge whether the distance of the coal pile reaches the threshold and obtain the decision result , and uses the neural network classifier trained based on image features to judge whether the distance of the coal pile reaches the threshold and obtain the decision result , and assigns different weights according to the importance of distance features and image features when judging the distance, that is and , and uses the weighted voting method to fuse the two decision results and , so as to comprehensively judge whether there is a coal piling situation and the height of the coal pile, and can comprehensively and accurately monitor and analyze the coal piling situation in real time, and can perform real-time intelligent control on the coal piling situation, realizing accurate monitoring and timely response to the coal piling situation, providing a reliable safety guarantee for coal mine production. The data fusion and judgment module uploads the fusion and judgment data to the display layer 5 and the coal mine terminal server through the network communication layer 4, so that the staff can view the change of the coal level in real time without going to the site. When it is judged that the coal piling is abnormal, the abnormal signal is sent to the display layer 5 and the coal mine terminal server through the network communication layer 4, and the coal mine terminal server automatically controls the conveyor to stop.

[0045] When the data processing module judges that the coal piling situation reaches or exceeds the warning threshold, it sends a warning signal to the coal mine underground monitoring system through the communication module. After receiving the warning signal, the monitoring system displays information such as the location and severity of the coal piling on the display screen and issues an audible and visual alarm to remind the staff to handle it in time. At the same time, the sensor continuously uploads the coal piling detection data to the monitoring system through the communication module so that the staff can grasp the dynamic coal piling on the belt conveyor in real time.

[0046] The above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements will not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A mine-used intrinsically safe radar coal pile sensor with an echo processing module, characterized in that: It includes a radar monitoring layer (1), a machine vision monitoring layer (2), a data processing and control layer (3), a network communication layer (4), a display layer (5), an audible and visual alarm layer (6) and a power supply layer (7); The radar monitoring layer (1) uses millimeter wave radar technology to transmit radar signals to the coal pile area to be measured and receive radar echo signals reflected from the coal pile surface, filter, mix and digitally process the radar echo signals, extract frequency characteristic information and phase characteristic information of the echo signals, and obtain distance data of the coal pile area to be measured; The machine vision monitoring layer (2) uses an industrial camera to capture an image of the coal pile area, and performs denoising, defogging, grayscale, edge detection, three-frame difference method and binarization on the image to obtain a machine vision processing result; The data processing and control layer (3) normalizes the distance data of the measured coal pile area and the machine vision processing result data, and combines the pre-set threshold value to perform data fusion and comprehensive judgment to determine whether there is a coal pile and the height of the coal pile. When it is determined that the coal pile is abnormal, the abnormal signal is transmitted to the display layer (5) and the coal mine terminal server through the network communication layer (4), and the coal mine terminal server automatically controls the conveyor to stop.

2. The mine intrinsically safe radar coal pile sensor with echo processing module according to claim 1, characterized in that: The network communication layer (4) is used to receive data from each layer, for data transmission and aggregation, and transmits the data to the coal mine terminal server through a 4G / 5G or ZigBee wireless communication module. The display layer (5) receives the coal pile data from the data processing and control layer (3) and displays it through an LED display screen. The sound and light alarm layer (6) receives abnormal signals from the data processing and control layer (3) and issues a sound and light alarm prompt. The power supply layer (7) uses a converter to convert the external input intrinsically safe power supply into a stable DC voltage required by each layer, thereby supplying power to each layer of the mine intrinsically safe radar coal pile sensor.

3. The intrinsically safe radar coal pile sensor with echo processing module for mine use according to claim 2, characterized in that: The radar monitoring layer (1) comprises a radar transmitting module, a radar receiving module, a preliminary amplification module, an echo processing unit and a radar data processing module; The radar transmitting module uses millimeter wave radar technology to transmit radar signals to the coal pile area to be tested; The radar receiving module receives the radar echo signal reflected from the surface of the coal pile; The preliminary amplification module performs preliminary amplification on the radar echo signal through a low noise amplifier; The echo processing unit extracts the frequency characteristic information and phase characteristic information of the echo signal through filtering, mixing and digital signal processing to obtain the distance data of the measured coal pile area; The radar data processing module adopts a sliding average filtering algorithm to average a plurality of distance data of the measured coal pile area continuously collected by the echo processing unit to obtain a coal pile height value.

4. The mine intrinsically safe radar coal pile sensor with echo processing module according to claim 3, characterized in that: The echo processing unit includes a signal filtering module, a frequency mixing module and a digital signal processing module; The signal filtering module uses a bandpass filter to filter out high-frequency noise and low-frequency interference signals in the echo signal and retain the frequency components related to the target reflection signal; The mixing module mixes the filtered echo signal with the local oscillation signal to convert the high-frequency echo signal into an intermediate-frequency signal; The digital signal processing module performs spectrum analysis on the intermediate frequency signal, extracts frequency characteristic information and phase characteristic information of the echo signal, and obtains distance data of the measured coal pile area.

5. The mine intrinsically safe radar coal pile sensor with echo processing module according to claim 4, characterized in that: The digital signal processing module performs spectrum analysis on the intermediate frequency signal to extract frequency characteristic information and phase characteristic information of the echo signal, including the following steps: Sampling and quantization: Sampling and quantizing the intermediate frequency signal, converting the analog intermediate frequency signal into a digital signal. The sampled signal is quantized to convert the continuous analog amplitude into a discrete digital value. Fast Fourier Transform: Perform fast Fourier transform on the sampled and quantized digital signal. Through fast Fourier transform analysis, the spectrum characteristics of the echo signal, that is, the distribution of the signal at different frequencies, are obtained. By analyzing the spectrum, the frequency peak related to the coal pile reflection signal is found; Feature extraction and analysis: By analyzing the phase and amplitude characteristics of the echo signal, the frequency characteristic information and phase characteristic information of the echo signal are extracted to obtain the distance data of the measured coal pile area.

6. The mine intrinsically safe radar coal pile sensor with echo processing module according to claim 1, characterized in that: The machine vision monitoring layer (2) comprises an image acquisition module, an image denoising module, an image defogging enhancement processing module, a grayscale processing module, an edge detection module, a three-frame difference method processing module and a binarization processing module; The image acquisition module uses an industrial camera to capture images of the coal pile area; The image denoising module receives image data and reads video frames, processes the image using a wavelet denoising method, obtains wavelet coefficients of different directions and scales, estimates a reasonable threshold through VisuShrink threshold processing, sets the wavelet coefficients less than the threshold to 0, and retains them otherwise. Based on the calculation result of the threshold estimation, the processed wavelet coefficients are further inversely transformed to obtain a denoised coal pile image. The function expression of the VisuShrink threshold is as follows: ； Where: A represents the threshold, represents the standard deviation of image noise, B represents the product of image width and height, represents the median function, Represents the wavelet function at the minimum layer; The image defogging and enhancement processing module uses a dark channel image defogging and enhancement algorithm to perform defogging and enhancement processing on the denoised coal pile image to obtain a defogged image; The grayscale processing module grayscales the defogged image using a weighted average method, and the image grayscale value is expressed as: ； Where: represents the red component in the image, Represents the green component in the image, Represents the blue component in the image; The edge detection module uses the Canny operator to perform edge detection on the grayscale processed image and calculates the direction of each pixel in the image. and gradient , perform non-maximum suppression on the gradient amplitude to obtain the edge intensity map, accurately locate the edge, and then use the double threshold algorithm to refine and connect the edge to obtain the edge image: ； Where: Represents the horizontal gradient of the pixel. Represents the gradient of the pixel in the vertical direction; The three-frame difference processing module uses a difference algorithm to analyze adjacent frame images, compares the grayscale values ​​of pixels at the same position, calculates their differences, and combines adjacent pixels for connectivity analysis. If the absolute value of the difference exceeds a set threshold, these changes are determined to be moving targets and a complete moving target outline is formed to achieve the extraction of moving targets. The binarization processing module further converts the image into an image with only black and white values ​​to obtain a machine vision processing result.

7. The mine intrinsically safe radar coal pile sensor with echo processing module according to claim 2, characterized in that: The data processing and control layer (3) includes a normalization module, a data fusion and judgment module and a self-checking module; The normalization module receives the distance data of the coal pile area to be measured output by the radar monitoring layer (1) and the machine vision processing result data output by the machine vision monitoring layer (2), and performs normalization processing on them; The data fusion and judgment module fuses and judges the two sets of data after normalization to determine whether there is a coal pile and the height of the coal pile; The self-check module performs a system self-check after the sensor is powered on, checking whether the hardware connections of each module are normal and whether the software program can run normally. If an abnormality is found during the self-check process, the data is uploaded to the display layer (5) through the network communication layer (4) for display.

8. The mine intrinsically safe radar coal pile sensor with echo processing module according to claim 7, characterized in that: The data fusion and judgment module fuses and judges the two normalized data, including the following steps: B1. Performing original data synchronization of the distance data features of the measured coal pile area and the data features of the machine vision processing results, i.e., spatial synchronization and temporal synchronization; B2. Combine the pre-set thresholds to make independent judgments on the two sets of synchronized data, using a support vector machine classifier trained based on distance features. Determine whether the distance to the coal pile reaches the threshold and obtain the decision result , using a neural network classifier trained on image features Determine whether the distance to the coal pile reaches the threshold and obtain the decision result ; B3, assign different weights according to the importance of distance features and image features in judging distance, that is and , using weighted voting to compare the two decision results and To perform the fusion: ； Where: and It is a binary classification result. If the distance to the coal pile reaches the threshold, it is 1, which means that the coal pile situation and the coal pile height exist. If it does not reach the threshold, it is 0, which means that the coal pile situation does not exist. B4. Model training and evaluation: Use the fused feature data to train the classification model, which can be a logistic regression model, a random forest model or a deep learning model. Use the K-fold cross-validation method to evaluate the trained model. Divide the data set into K parts, use one of them as the test set each time, and use the remaining K-1 parts as the training set. Repeat K times, calculate the evaluation indicators each time, namely, accuracy, recall rate and F1 value, and take the average value as the final evaluation result; B5, the data fusion and judgment module uploads the fusion and judgment data to the display layer (5) and the coal mine terminal server through the network communication layer (4); B6. When the data fusion and judgment module detects an abnormality in the coal pile, the abnormal signal is transmitted to the display layer (5) and the coal mine terminal server through the network communication layer (4), and the coal mine terminal server automatically controls the conveyor to stop.

Images