Coal transportation monitoring device and abnormality detection method

Abstract

The present invention relates to the field of data processing technology, and in particular to a coal conveying monitoring device and an abnormality detection method, comprising an image acquisition system, a target detection system, a comparison system and a fault warning system, wherein the image acquisition system performs image acquisition on a coal conveying belt during operation to obtain a belt image set; the target detection system sequentially inputs the belt image set into a target detection model for training in a shooting order to obtain a belt dynamic data set; the comparison system compares the belt dynamic data set with the belt static data to obtain comparison data, and combines the comparison data with the belt vibration data collected by a vibration sensor to obtain the belt vibration condition; the fault warning system analyzes the belt vibration condition to obtain an analysis result, and issues a warning reminder based on the analysis result, thereby solving the problem that the existing coal conveying monitoring device has a poor detection effect on the vibration of the coal conveying belt.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a coal conveying monitoring device and an abnormality detection method. Background Art

[0002] Currently, coal conveying equipment is mainly belt conveyors. When a belt conveyor fails, it will vibrate. When the vibration amplitude is too small, it is not easy for workers to detect it with the naked eye. Therefore, it is necessary to monitor the belt conveyor that transports the coal flow through a coal conveying monitoring device.

[0003] At present, the prior art discloses a coal conveying monitoring device including a terminal display system, a vibration sensor, a single-chip microcomputer and a processor. The vibration sensor is installed on a mounting frame of a belt conveyor to monitor the coal conveyor belt of the belt conveyor. The processor processes the monitoring data and displays it on the terminal display system.

[0004] With the above method, when the coal conveyor belt is conveying coal flow, the collision of the coal flow with the vibration sensor will cause the vibration sensor to shift, thereby reducing the detection effect of the vibration of the coal conveyor belt. Summary of the invention

[0005] The object of the present invention is to provide a coal conveying monitoring device and an abnormality detection method, aiming to solve the problem that the existing coal conveying monitoring device has a poor detection effect on the vibration of the coal conveying belt.

[0006] To achieve the above-mentioned object, in a first aspect, the present invention provides a coal transportation monitoring device, comprising an image acquisition system, a target detection system, a comparison system and a fault warning system, wherein the image acquisition system, the target detection system, the comparison system and the fault warning system are connected in sequence;

[0007] The image acquisition system is used to acquire images of the coal conveyor belt during operation to obtain a belt image set;

[0008] The target detection system is used to input the belt image set into the target detection model in sequence according to the shooting order for training, so as to obtain a belt dynamic data set;

[0009] The comparison system is used to compare the belt dynamic data set with the belt static data to obtain comparison data, and combine the comparison data with the belt vibration data collected by the vibration sensor to obtain the belt vibration condition;

[0010] The fault warning system is used to analyze the vibration of the belt, obtain analysis results, and issue a warning reminder based on the analysis results.

[0011] Wherein, the image acquisition system comprises a distance measurement module, a parameter adjustment module, an acquisition module and a preprocessing module, and the distance measurement module, the parameter adjustment module, the acquisition module and the preprocessing module are connected in sequence;

[0012] The distance measuring module is used to measure the distance between the collection end of the collection module and the coal conveying belt;

[0013] The parameter adjustment module adjusts the resolution and the number of shots of the acquisition module based on the spacing;

[0014] The acquisition module is used to continuously shoot the coal conveyor belt after adjustment to obtain a continuously shot image set;

[0015] The preprocessing module is used to preprocess the continuously shot image set to obtain a belt image set.

[0016] Wherein, the target detection system comprises a construction module, a training module and an input module, and the construction module, the training module and the input module are connected in sequence;

[0017] The building block is used to build a deep learning neural network;

[0018] The training module is used to train and verify the deep learning neural network using a public data set to obtain a target detection model;

[0019] The input module is used to input the belt image set into the target detection model in sequence according to the shooting order for training, so as to obtain a belt dynamic data set.

[0020] Wherein, the training module includes an acquisition submodule, a labeling submodule, a division submodule and a training submodule, and the acquisition submodule, the labeling submodule, the division submodule and the training submodule are connected in sequence;

[0021] The acquisition submodule is used to acquire a public data set;

[0022] The annotation submodule is used to perform target annotation on the public data set to obtain an annotated data set;

[0023] The division submodule is used to divide the labeled data set into a training set and a validation set;

[0024] The training submodule is used to train the deep learning neural network using the training set to obtain a pre-trained model, and then use the verification set to verify the pre-trained model. If the verification passes, a target detection module is obtained.

[0025] Wherein, the acquisition submodule includes an acquisition unit and a filtering unit, and the acquisition unit is connected to the filtering unit;

[0026] The acquisition unit is used to crawl the target image on the website to obtain a crawled data set;

[0027] The filtering unit is used to filter the crawled data set to obtain a public data set.

[0028] Wherein, the comparison system comprises a position selection module, a drawing module and a comparison module, and the position selection module, the drawing module and the comparison module are connected in sequence;

[0029] The position selection module is used to select a reference point on the static data of the belt, and select a corresponding comparison position point on each picture of the dynamic data set of the belt based on the reference point;

[0030] The drawing module is used to draw a position fluctuation diagram using each of the compared position points;

[0031] The comparison module is used to compare the position fluctuation diagram with the reference point to obtain the belt vibration condition.

[0032] Among them, the acquisition module includes a pan-tilt head, a camera, a shell, a reset component and a cleaning component. The pan-tilt head is fixedly connected to the shell and is located on the top of the shell. The camera is fixedly connected to the output end of the pan-tilt head. The reset component is arranged in the shell and contacts the bottom of the camera. The cleaning component is arranged on one side of the reset component.

[0033] In a second aspect, the present invention provides a method for detecting an abnormality of a coal conveying monitoring device, comprising the following steps:

[0034] The image acquisition system is used to acquire images of the coal conveyor belt during operation to obtain a belt image set;

[0035] The belt image set is input into the target detection model in the order of shooting through the target detection system for training, and the belt dynamic data set is obtained;

[0036] The belt dynamic data set is compared with the belt static data by a comparison system to obtain comparison data, and the comparison data is combined with the belt vibration data collected by the vibration sensor to obtain the belt vibration condition;

[0037] The fault warning system analyzes the belt vibration condition, obtains the analysis results, and issues warning reminders based on the analysis results.

[0038] A coal conveying monitoring device of the present invention firstly collects images of a coal conveying belt during operation through the image acquisition system to obtain a belt image set; then, the belt image set is input into a target detection model for training in the order of shooting through the target detection system to obtain a belt dynamic data set; then, the comparison system compares the belt dynamic data set with the belt static data to obtain comparison data, and combines the comparison data with the belt vibration data collected by the vibration sensor to obtain the belt vibration condition; finally, the fault warning system analyzes the belt vibration condition to obtain an analysis result, and issues a warning reminder based on the analysis result. The present invention performs target detection and post-analysis after picture acquisition through the image acquisition system, and the coal conveying belt will not cause any influence on the operation of the image acquisition system during the coal flow transportation process, thereby improving the detection effect of the coal conveying belt vibration, and solving the problem that the existing coal conveying monitoring device has a poor detection effect on the coal conveying belt vibration. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0040] Figure 1 It is a structural schematic diagram of a coal conveying monitoring device provided by the present invention.

[0041] Figure 2 It is a structural diagram of the training module.

[0042] Figure 3 It is a schematic diagram of the structure of the acquisition submodule.

[0043] Figure 4 It is a structural diagram of the acquisition module.

[0044] Figure 5 It is a cross-sectional view of the acquisition module.

[0045] Figure 6 The present invention provides a flow chart of a method for detecting anomalies in a coal conveying monitoring device.

[0046] 1-image acquisition system, 2-target detection system, 3-comparison system, 4-fault warning system, 5-distance measurement module, 6-parameter adjustment module, 7-acquisition module, 8-preprocessing module, 9-construction module, 10-training module, 11-input module, 12-acquisition submodule, 13-labeling submodule, 14-division submodule, 15-training submodule, 16-acquisition unit, 17-filtering unit, 18-position selection module, 19-drawing module, 20-comparison module, 21-pan head, 22-camera, 23-housing, 24-reset component, 25-cleaning component, 26-reset spring, 27-inclined table, 28-pull rod, 29-motor, 30-rotating shaft, 31-elastic rod, 32-cleaning brush, 33-dust collection box, 34-slide rail, 35-return spring, 36-slide plate. DETAILED DESCRIPTION

[0047] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0048] See also Figures 1 to 5 The present invention provides a coal transportation monitoring device, comprising an image acquisition system 1, a target detection system 2, a comparison system 3 and a fault warning system 4, wherein the image acquisition system 1, the target detection system 2, the comparison system 3 and the fault warning system 4 are connected in sequence;

[0049] The image acquisition system 1 is used to acquire images of the coal conveyor belt during operation to obtain a belt image set;

[0050] The target detection system 2 is used to input the belt image set into the target detection model in sequence according to the shooting order for training, so as to obtain a belt dynamic data set;

[0051] The comparison system 3 is used to compare the belt dynamic data set with the belt static data to obtain comparison data, and combine the comparison data with the belt vibration data collected by the vibration sensor to obtain the belt vibration situation;

[0052] The fault warning system 4 is used to analyze the vibration of the belt, obtain analysis results, and issue warning reminders based on the analysis results.

[0053] Specifically, first, the image acquisition system 1 is used to acquire images of the coal conveyor belt during operation to obtain a belt image set; then, the target detection system 2 inputs the belt image set into the target detection model in the order of shooting for training to obtain a belt dynamic data set; then, the comparison system 3 compares the belt dynamic data set with the belt static data to obtain comparison data, and combines the comparison data with the belt vibration data acquired by the vibration sensor to obtain the belt vibration condition; finally, the fault warning system 4 analyzes the belt vibration condition to obtain an analysis result, and issues a warning reminder based on the analysis result. The present invention performs target detection and post-analysis after image acquisition by the image acquisition system 1, and the coal conveyor belt does not cause any influence on the operation of the image acquisition system 1 during the coal flow transportation process, thereby improving the detection effect of the coal conveyor belt vibration, and solving the problem that the existing coal conveyor monitoring device has poor detection effect on the coal conveyor belt vibration.

[0054] Further, the image acquisition system 1 includes a distance measurement module 5, a parameter adjustment module 6, an acquisition module 7 and a preprocessing module 8, and the distance measurement module 5, the parameter adjustment module 6, the acquisition module 7 and the preprocessing module 8 are connected in sequence;

[0055] The distance measuring module 5 is used to measure the distance between the collection end of the collection module 7 and the coal conveying belt;

[0056] The parameter adjustment module 6 adjusts the resolution and the number of shots of the acquisition module 7 based on the spacing;

[0057] The acquisition module 7 is used to continuously shoot the coal conveyor belt after adjustment to obtain a continuously shot image set;

[0058] The preprocessing module 8 is used to preprocess the continuously shot image set to obtain a belt image set.

[0059] Specifically, first, the distance measuring module 5 measures the distance between the collecting end of the collecting module 7 and the coal conveyor belt. The distance measuring module 5 can be a processing unit, a photosensitive sensor and a light source. The light source is arranged on the coal conveyor belt, and the photosensitive sensor is arranged on the collecting module 7. After the photosensitive sensor converts the received light signal of the light source into an electrical signal, the processing unit processes the electrical signal to obtain the distance between the collecting end of the collecting module 7 and the coal conveyor belt; then, the parameter adjustment module 6 adjusts the resolution and the number of shots of the collecting module 7 based on the distance; then, the collecting module 7 continuously shoots the coal conveyor belt after adjustment to obtain a continuously shot image set; finally, the preprocessing module 8 preprocesses the continuously shot image set, adjusts the exposure, contrast image and shadow of each image in the continuously shot image set, and obtains a belt image set.

[0060] Further, the target detection system 2 includes a construction module 9, a training module 10 and an input module 11, and the construction module 9, the training module 10 and the input module 11 are connected in sequence;

[0061] The construction module 9 is used to construct a deep learning neural network;

[0062] The training module 10 is used to train and verify the deep learning neural network using a public data set to obtain a target detection model;

[0063] The input module 11 is used to input the belt image set into the target detection model in sequence according to the shooting order for training, so as to obtain a belt dynamic data set.

[0064] The training module 10 includes an acquisition submodule 12, a labeling submodule 13, a division submodule 14 and a training submodule 15, wherein the acquisition submodule 12, the labeling submodule 13, the division submodule 14 and the training submodule 15 are connected in sequence;

[0065] The acquisition submodule 12 is used to acquire a public data set;

[0066] The annotation submodule 13 is used to perform target annotation on the public data set to obtain an annotated data set;

[0067] The division submodule 14 is used to divide the labeled data set into a training set and a validation set;

[0068] The training submodule 15 is used to train the deep learning neural network using the training set to obtain a pre-trained model, and then use the verification set to verify the pre-trained model. If the verification passes, a target detection module is obtained.

[0069] The acquisition submodule 12 includes an acquisition unit 16 and a filtering unit 17, and the acquisition unit 16 is connected to the filtering unit 17;

[0070] The acquisition unit 16 is used to crawl the target image on the website to obtain a crawled data set;

[0071] The filtering unit 17 is used to filter the crawled data set to obtain a public data set.

[0072] Specifically, first, the construction module 9 constructs a deep learning neural network; secondly, the acquisition unit 16 crawls the target image on the website to obtain a crawled data set; the filtering unit 17 filters the crawled data set to obtain a public data set; then, the annotation submodule 13 performs target annotation on the public data set to obtain a labeled data set; the division submodule 14 divides the labeled data set to obtain a training set and a verification set; then, the training submodule 15 uses the training set to train the deep learning neural network to obtain a pre-trained model, and then uses the verification set to verify the pre-trained model. If the verification passes, a target detection module is obtained. Finally, the input module 11 inputs the belt image set into the target detection model in the order of shooting for training to obtain a belt dynamic data set.

[0073] Further, the comparison system 3 includes a position selection module 18, a drawing module 19 and a comparison module 20, and the position selection module 18, the drawing module 19 and the comparison module 20 are connected in sequence;

[0074] The position selection module 18 is used to select a reference point on the belt static data, and select a corresponding comparison position point on each picture of the belt dynamic data set based on the reference point;

[0075] The drawing module 19 is used to draw a position fluctuation diagram using each of the compared position points;

[0076] The comparison module 20 is used to compare the position fluctuation diagram with the reference point to obtain comparison data, and combine the comparison data with the belt vibration data collected by the vibration sensor to obtain the belt vibration condition.

[0077] Specifically, first, the position selection module 18 selects a reference point on the static data of the belt, and selects a corresponding comparison position point on each picture of the dynamic data set of the belt based on the reference point, the static data of the belt is the data of the coal conveyor belt when it is not working, that is, the data when it is not vibrating, and the reference point and the comparison position point are the same position on the coal conveyor belt; then, the drawing module 19 uses each of the comparison position points to draw a position fluctuation graph; finally, the comparison module 20 compares the position fluctuation graph with the reference point to obtain comparison data, and combines the comparison data with the belt vibration data collected by the vibration sensor to obtain the belt vibration condition.

[0078] Furthermore, the acquisition module 7 includes a pan-tilt head 21, a camera 22, a housing 23, a reset component 24 and a cleaning component 25. The pan-tilt head 21 is fixedly connected to the housing 23 and is located on the top of the housing 23. The camera 22 is fixedly connected to the output end of the pan-tilt head 21. The reset component 24 is arranged in the housing 23 and contacts the bottom of the camera 22. The cleaning component 25 is arranged on one side of the reset component 24.

[0079] Specifically, the pan / tilt head 21 is installed on a wall and provides installation conditions for the camera 22 and the housing 23. The pan / tilt head 21 is used to drive the camera 22 to rotate. When the camera 22 rotates, the reset component 24 is squeezed to slide into the housing 23. When the reset component 24 slides into the housing 23, it drives the cleaning component 25 to move downward until the acquisition end of the camera 22 contacts the cleaning component 25. The cleaning component 25 works to clean the acquisition end of the camera 22 to prevent impurities on the acquisition end from affecting the clarity of the picture collected by the camera 22. A cooling fan is arranged on the side of the camera 22 away from the acquisition end. The housing 23 has a heat dissipation port and an air inlet. The cooling fan discharges the hot air in the housing 23 through the heat dissipation port, and cold air enters the housing 23 from the air inlet to cool the camera 22.

[0080] Furthermore, the reset assembly 24 includes a reset spring 26, a ramp 27 and a pull rod 28. The reset spring 26 is fixedly connected to the housing 23 and is located on the inner wall of the housing 23. The ramp 27 is fixedly connected to the reset spring 26 and is slidably connected to the housing 23. The pull rod 28 is fixedly connected to the ramp 27 and is located on a side away from the reset spring 26.

[0081] Specifically, when the cleaning component 25 cleans the acquisition end of the camera 22, the camera 22 presses the inclined platform 27 into the inner wall of the outer shell 23 and squeezes the reset spring 26. After cleaning, the pan-tilt head 21 reverses and drives the camera 22 to reset. At this time, the inclined platform 27 is not subject to force, and the reset spring 26 resets and pushes the inclined platform 27 to slide out of the outer shell 23, and drives the pull rod 28 to move. The pull rod 28 drives the cleaning component 25 to move upward and staggered with the cooling fan on the back of the camera 22 to avoid the cleaning component 25 affecting the operation of the cooling fan.

[0082] Furthermore, the cleaning assembly 25 includes a motor 29, a rotating shaft 30, an elastic rod 31, a cleaning brush 32 and a dust box 33. The motor 29 is fixedly connected to the pull rod 28 and is located on the outer wall of the pull rod 28. The rotating shaft 30 is fixedly connected to the output end of the motor 29. The elastic rod 31 is fixedly connected to the rotating shaft 30 and is located on a side away from the motor 29. The cleaning brush 32 is fixedly connected to the elastic rod 31 and is located on a side away from the rotating shaft 30. The dust box 33 is fixedly connected to the outer shell 23 and passes through the outer shell 23.

[0083] Specifically, the motor 29 drives the rotating shaft 30 to drive the elastic rod 31 to rotate. When the elastic rod 31 rotates, it drives the cleaning brush 32 to rotate to clean the collection end of the camera 22. At the same time, the elastic rod 31 pushes the cleaning brush 32 toward one side of the camera 22, thereby reducing the gap between the cleaning brush 32 and the camera 22, thereby increasing the cleaning effect of the camera 22. The dust box 33 absorbs the impurities cleaned by the cleaning brush 32 and stores them.

[0084] Furthermore, the elastic rod 31 includes a slide rail 34, a return spring 35 and a slide plate 36. The slide rail 34 is fixedly connected to the rotating shaft 30 and is located on a side away from the motor 29. The return spring 35 is fixedly connected to the slide rail 34 and is located inside the slide rail 34. The slide plate 36 is fixedly connected to the return spring 35, slidably connected to the slide rail 34, and is located on the inner wall of the slide rail 34.

[0085] Specifically, the slide rail 34 provides installation conditions for the return spring 35 and the slide plate 36 , the return spring 35 pushes the slide plate 36 toward the side close to the camera 22 , and the slide plate 36 pushes the cleaning brush 32 to contact the collection end of the camera 22 .

[0086] See also Figure 6 In a second aspect, the present invention provides a method for detecting abnormality of a coal conveying monitoring device, comprising the following steps:

[0087] S1 collects images of the coal conveyor belt during operation through the image acquisition system 1 to obtain a belt image set;

[0088] Specifically, first, the distance measuring module 5 measures the distance between the collecting end of the collecting module 7 and the coal conveyor belt. The distance measuring module 5 can be a processing unit, a photosensitive sensor and a light source. The light source is arranged on the coal conveyor belt, and the photosensitive sensor is arranged on the collecting module 7. After the photosensitive sensor converts the received light signal of the light source into an electrical signal, the processing unit processes the electrical signal to obtain the distance between the collecting end of the collecting module 7 and the coal conveyor belt; then, the parameter adjustment module 6 adjusts the resolution and the number of shots of the collecting module 7 based on the distance; then, the collecting module 7 continuously shoots the coal conveyor belt after adjustment to obtain a continuously shot image set; finally, the preprocessing module 8 preprocesses the continuously shot image set, adjusts the exposure, contrast image and shadow of each image in the continuously shot image set, and obtains a belt image set.

[0089] S2 inputs the belt image set into the target detection model in the order of shooting through the target detection system 2 for training, and obtains the belt dynamic data set;

[0090] Specifically, first, the construction module 9 constructs a deep learning neural network; secondly, the acquisition unit 16 crawls the target image on the website to obtain a crawled data set; the filtering unit 17 filters the crawled data set to obtain a public data set; then, the annotation submodule 13 performs target annotation on the public data set to obtain a labeled data set; the division submodule 14 divides the labeled data set to obtain a training set and a verification set; then, the training submodule 15 uses the training set to train the deep learning neural network to obtain a pre-trained model, and then uses the verification set to verify the pre-trained model. If the verification passes, a target detection module is obtained. Finally, the input module 11 inputs the belt image set into the target detection model in the order of shooting for training to obtain a belt dynamic data set.

[0091] S3 compares the belt dynamic data set with the belt static data through the comparison system 3 to obtain comparison data, and combines the comparison data with the belt vibration data collected by the vibration sensor to obtain the belt vibration condition;

[0092] Specifically, first, the position selection module 18 selects a reference point on the static data of the belt, and based on the reference point, selects a corresponding comparison position point on each picture of the dynamic data set of the belt, the static data of the belt is the data of the coal conveyor belt when it is not working, that is, the data when it is not vibrating, and the reference point and the comparison position point are the same position on the coal conveyor belt; then, the drawing module 19 uses each of the comparison position points to draw a position fluctuation diagram; finally, the comparison module 20 compares the position fluctuation diagram with the reference point to obtain the vibration condition of the belt.

[0093] S4 analyzes the belt vibration condition through the fault warning system 4, obtains the analysis result, and issues a warning reminder based on the analysis result.

[0094] Specifically, the fault warning system 4 analyzes the belt vibration condition to classify the fault condition and obtain the analysis result, and based on the analysis result, it is convenient to provide early warning in the later stage. At the same time, the belt vibration condition and the belt dynamic data are uploaded to the host computer for storage.

[0095] What is disclosed above is only a preferred embodiment of a coal conveying monitoring device and anomaly detection method of the present invention. Of course, this cannot be used to limit the scope of rights of the present invention. Ordinary technicians in this field can understand that all or part of the processes of the above embodiments and equivalent changes made according to the claims of the present invention still fall within the scope of the invention.

Claims

1. A coal conveying monitoring device, characterized in that: It includes an image acquisition system, a target detection system, a comparison system and a fault warning system, wherein the image acquisition system, the target detection system, the comparison system and the fault warning system are connected in sequence; The image acquisition system is used to acquire images of the coal conveyor belt during operation to obtain a belt image set; The target detection system is used to input the belt image set into the target detection model in sequence according to the shooting order for training, so as to obtain a belt dynamic data set; The comparison system is used to compare the belt dynamic data set with the belt static data to obtain comparison data, and combine the comparison data with the belt vibration data collected by the vibration sensor to obtain the belt vibration condition; The fault warning system is used to analyze the vibration of the belt, obtain analysis results, and issue warning reminders based on the analysis results; The image acquisition system comprises a distance measurement module, a parameter adjustment module, an acquisition module and a preprocessing module, wherein the distance measurement module, the parameter adjustment module, the acquisition module and the preprocessing module are connected in sequence; The distance measuring module is used to measure the distance between the collection end of the collection module and the coal conveying belt; The parameter adjustment module adjusts the resolution and the number of shots of the acquisition module based on the spacing; The acquisition module is used to continuously shoot the coal conveyor belt after adjustment to obtain a continuously shot image set; The preprocessing module is used to preprocess the continuously shot image set to obtain a belt image set; The acquisition module includes a pan-tilt head, a camera, a housing, a reset component and a cleaning component. The pan-tilt head is fixedly connected to the housing and is located on the top of the housing. The camera is fixedly connected to the output end of the pan-tilt head. The reset component is arranged in the housing and contacts the bottom of the camera. The cleaning component is arranged on one side of the reset component. The reset assembly includes a reset spring, a ramp and a pull rod, wherein the reset spring is fixedly connected to the housing and is located on the inner wall of the housing, the ramp is fixedly connected to the reset spring and is slidably connected to the housing, and the pull rod is fixedly connected to the ramp and is located on a side away from the reset spring; The cleaning assembly includes a motor, a rotating shaft, an elastic rod, a cleaning brush and a dust collection box, wherein the motor is fixedly connected to the pull rod and is located on the outer side wall of the pull rod, the rotating shaft is fixedly connected to the output end of the motor, the elastic rod is fixedly connected to the rotating shaft and is located on a side away from the motor, the cleaning brush is fixedly connected to the elastic rod and is located on a side away from the rotating shaft, and the dust collection box is fixedly connected to the shell and passes through the shell; The elastic rod includes a slide rail, a return spring and a slide plate. The slide rail is fixedly connected to the rotating shaft and is located on a side away from the motor. The return spring is fixedly connected to the slide rail and is located inside the slide rail. The slide plate is fixedly connected to the return spring, slidably connected to the slide rail, and is located on the inner wall of the slide rail.

2. The coal transport monitoring device according to claim 1, characterized in that: The target detection system comprises a construction module, a training module and an input module, wherein the construction module, the training module and the input module are connected in sequence; The building block is used to build a deep learning neural network; The training module is used to train and verify the deep learning neural network using a public data set to obtain a target detection model; The input module is used to input the belt image set into the target detection model in sequence according to the shooting order for training, so as to obtain a belt dynamic data set.

3. The coal transport monitoring device according to claim 2, characterized in that: The training module includes an acquisition submodule, a labeling submodule, a division submodule and a training submodule, and the acquisition submodule, the labeling submodule, the division submodule and the training submodule are connected in sequence; The acquisition submodule is used to acquire a public data set; The annotation submodule is used to perform target annotation on the public data set to obtain an annotated data set; The division submodule is used to divide the labeled data set into a training set and a validation set; The training submodule is used to train the deep learning neural network using the training set to obtain a pre-trained model, and then use the verification set to verify the pre-trained model. If the verification passes, a target detection module is obtained.

4. The coal transport monitoring device according to claim 3, characterized in that: The acquisition submodule includes an acquisition unit and a filtering unit, and the acquisition unit is connected to the filtering unit; The acquisition unit is used to crawl the target image on the website to obtain a crawled data set; The filtering unit is used to filter the crawled data set to obtain a public data set.

5. The coal transport monitoring device according to claim 4, characterized in that: The comparison system comprises a position selection module, a drawing module and a comparison module, wherein the position selection module, the drawing module and the comparison module are connected in sequence; The position selection module is used to select a reference point on the static data of the belt, and select a corresponding comparison position point on each picture of the dynamic data set of the belt based on the reference point; The drawing module is used to draw a position fluctuation diagram using each of the compared position points; The comparison module is used to compare the position fluctuation diagram with the reference point to obtain the belt vibration condition.

6. A method for detecting abnormality of a coal conveying monitoring device, applied to the coal conveying monitoring device according to claims 1-5, characterized in that: The following steps are involved: The image acquisition system is used to acquire images of the coal conveyor belt during operation to obtain a belt image set; The belt image set is input into the target detection model in the order of shooting through the target detection system for training, and the belt dynamic data set is obtained; The belt dynamic data set is compared with the belt static data through a comparison system to obtain comparison data, and the comparison data is combined with the belt vibration data collected by the vibration sensor to obtain the belt vibration condition; The fault warning system analyzes the belt vibration condition, obtains the analysis results, and issues warning reminders based on the analysis results.

Images