Dust particle size and concentration detection device and method based on image analysis

Abstract

The invention discloses a dust particle size and concentration detection device and method based on image analysis, and the device comprises a box body, a punching plate is fixed at the middle part in the box body, a dust sample collection mechanism is arranged in an upper cavity of the box body, a dust image collection mechanism is arranged at the lower side of the dust sample collection mechanism, and an upper computer is arranged on the side wall of the upper cavity of the box body. The upper computer comprises a transmission module, an image processing module and a display module, the transmission module is in communication connection with the image processing module, and the output end of the image processing module is connected with the input end of the display module. Dust sample images are collected through the industrial camera, the collected sample images are transmitted to the image processing module of the upper computer to be processed, and the image processing module analyzes dust concentration and dust particle size distribution data. Workers can more comprehensively know the dust condition in the coal mine underground working environment and take measures in time for treatment, and the health guarantee of the coal mine underground workers is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of coal mine dust detection, and in particular to a device and method for detecting dust particle size and concentration based on image analysis. Background Art

[0002] In my country, coal is the main energy source, accounting for more than 70% of primary energy production and consumption. However, with the increase in coal exploration and mining efforts and the rapid improvement of tunneling technology and coal mining efficiency, the coal mine comprehensive excavation face, as the main link in the generation of dust from underground operations, has caused increasingly serious dust accumulation pollution problems. Dust control has become an important issue in coal mine safety production.

[0003] Dust detection is a very important part of coal mine dust control. First, excessive dust concentration may cause explosion accidents, seriously threatening the safe production of coal mines; second, long-term exposure to high-concentration dust will have a serious impact on the health of workers and increase the risk of diseases such as pneumoconiosis; therefore, dust detection is a necessary means to ensure the safe production of coal mines and protect the health of workers; however, there are many problems with existing mine dust concentration detection equipment. Most of the existing dust concentration monitoring equipment at home and abroad are based on weighing, light scattering, and radiography. The weighing method is used for manual measurement, cannot be detected online, and the operation is cumbersome; the light scattering method is used for dry flue gas or high concentration occasions, and the optical system is easily contaminated and maintenance is troublesome; the radiography ... The method is rarely used for online monitoring due to radiation and other reasons; moreover, the traditional detection means have a low degree of integration and cannot capture the dynamic changes of two parameters in real time. The method of separate sampling and measurement of different parameters is affected by transient airflow disturbances and inconsistent measurement time, and it is difficult to effectively characterize the physical parameters of the same dust group; at the same time, the lag of the detection process makes it difficult to effectively take prevention and control measures, which limits the dust control effect, and the dust concentration detection equipment can only measure the dust mass concentration, and cannot give the particle size distribution range at the same time; these problems lead to poor dust control and control effects, and it is impossible to obtain dust concentration change data and the distribution law of different dust particle size ranges (such as respirable dust) in time, and it is impossible to take timely measures for control. Summary of the Invention

[0004] The purpose of the present invention is to overcome the deficiencies in the above-mentioned prior art and provide a dust particle size and concentration detection device and method based on image analysis, including an industrial camera and a host computer. The industrial camera collects dust sample images and transmits the collected sample images to the image processing module of the host computer for processing. The image processing module in the present invention is established after model training of the initial image processing module. The image processing module analyzes the dust concentration and dust particle size distribution data based on the dust sample image. The staff can have a more comprehensive understanding of the dust situation in the underground working environment of the coal mine based on the dust concentration and dust particle size distribution data, and take timely measures to control it, thereby improving the health protection of the underground coal mine workers.

[0005] To achieve the above-mentioned purpose, the technical solution adopted by the present invention is: a dust particle size and concentration detection device based on image analysis, including a box body, a punching plate is fixed in the middle of the box body, the punching plate divides the inner cavity of the box body into an upper cavity and a lower cavity, a dust sample collection mechanism for collecting dust samples is provided in the upper cavity of the box body, a dust image collection mechanism for collecting images of dust samples to obtain dust sample images is provided on the lower side of the dust sample collection mechanism, a host computer is provided on the side wall of the upper cavity of the box body, the host computer includes a transmission module, an image processing module and a display module, the transmission module The block is communicatively connected to the image processing module and the transmission module is used to send and receive data. The transmission module receives the dust sample image transmitted by the dust image acquisition mechanism and transmits the dust sample image to the image processing module. The image processing module performs image analysis on the received dust sample image to obtain dust concentration and dust particle size distribution data. The output end of the image processing module is connected to the input end of the display module. The image processing module transmits the dust concentration and dust particle size distribution data to the display module for display. A dust discharge mechanism for discharging dust in the dust sample acquisition mechanism is provided in the lower cavity of the box.

[0006] Preferably, the host computer further includes a plurality of buttons, and the buttons on the host computer can be operated to issue instructions to control the start or stop of the dust sample collection mechanism, the dust image collection mechanism and the dust discharge mechanism.

[0007] Preferably, the image processing module includes a preprocessing model, a feature extraction model and a neural network model, and the neural network model is trained according to the following steps:

[0008] Step S1: Build an experimental platform and obtain a dust sample image set and label dataset based on the experimental platform;

[0009] Step S2: preprocessing the dust sample image set using the preprocessing model, extracting features from the preprocessed image using the feature extraction model, and performing correlation analysis and redundancy elimination on the feature parameters and the total dust concentration to obtain a feature set for input into the neural network model;

[0010] Step S3: The neural network model is trained according to the feature set and label data set of the input neural network model to obtain a trained neural network model.

[0011] Preferably, a dust channel is also provided on the upper part of the upper cavity of the box body, and the dust channel includes an intake fan, a ventilation duct, a ventilation port and an air outlet, and the ventilation port and the air outlet are respectively arranged on the left and right sides of the box body, and the intake fan is installed on the ventilation port, and the ventilation duct is installed between the ventilation port and the air outlet, and the left and right end portions of the ventilation duct are respectively connected to the ventilation port and the air outlet, and a first mounting port and a second mounting port are respectively provided on the upper and lower sides of the ventilation duct, and the first mounting port is located on the upper side of the second mounting port, and a ventilation duct fixing bracket is fixedly connected to the lower side of the ventilation duct, and the ventilation duct fixing bracket is fixedly connected to the punching plate at one end away from the ventilation duct.

[0012] Preferably, the dust sample collection mechanism includes a glass slide rack, a slide rail, a glass slide and a cover glass, the number of the slide rails is two, the two slide rails are respectively fixed on the front and rear sides of the top of the punching plate, each slide rail is provided with an electromagnet fixing frame that slides along the slide rail, each electromagnet fixing frame is installed with an electromagnet, the two electromagnets are respectively located on the front and rear sides of the ventilation pipe, the glass slide rack is a T-shaped structure, the vertical end portion of the glass slide rack is connected to the cover glass so that the cover glass is located in the first mounting port, the horizontal end portion of the glass slide rack is respectively provided with a first sleeve ring and a second sleeve ring, the first sleeve ring and the second sleeve ring are respectively sleeved on the electromagnets located on the front and rear sides of the ventilation pipe, the electromagnet drives the glass slide rack to move along the electromagnet when energized, and the glass slide is fixed in the second mounting port.

[0013] Preferably, the dust image acquisition mechanism includes a light source, a light source bracket, an industrial camera and a camera bracket, the light source bracket is a hollow bracket and the top and bottom of the bracket are open, the light source is a strip light source and is installed at the bottom opening of the light source bracket, the top of the light source bracket is installed at the second installation port and is fixedly connected to the wall of the ventilation pipe, the industrial camera is located in the light source bracket and the industrial camera is installed on the camera bracket, the camera bracket is a square support bracket, the bottom of the square support bracket is fixedly connected to the punching plate, the industrial camera is installed on the top of the square support bracket and the industrial camera is 50 mm away from the glass slide.

[0014] Preferably, the dust discharge mechanism includes a pressure gauge, an air pump, an air storage tank and a nozzle cover, the pressure gauge is installed on the upper side of the air pump for detecting the gas pressure in the air storage tank, the air pump is installed on the upper side of the air storage tank, an air intake hose is also installed on the air pump, the air intake hose is connected to the air storage tank at one end away from the air pump, the air pump transports gas into the air storage tank through the intake hose, an exhaust hose is also installed on the top of the air storage tank, the exhaust hose is provided with a manual valve near the end of the air storage tank, the nozzle cover is located at the bottom of the ventilation pipe and is fixedly connected to the ventilation pipe by bolts, a high-pressure nozzle and a flip baffle are installed in the nozzle cover, the flip baffle is installed on the upper side of the high-pressure nozzle, the left end of the flip baffle is rotatably connected to the inner side wall of the nozzle cover, and an electromagnetic valve is installed on the lower side of the high-pressure nozzle, and the electromagnetic valve is connected to the end of the exhaust hose away from the air storage tank.

[0015] Preferably, a containing frame is fixedly connected to the inner side wall of the lower cavity of the box body, a battery is arranged in the containing frame, a cover plate is provided on the outer side wall of the containing frame, the cover plate is detachably connected to the box body and the containing frame respectively, the battery is electrically connected to the air intake fan, air pump, electromagnet, light source, solenoid valve and upper machine respectively, and a first box door and a second box door are also provided on the front side of the box body, and the first box door is located on the upper side of the second box door.

[0016] The present invention also discloses a method for detecting dust particle size and concentration based on image analysis, comprising the following steps:

[0017] Step S1: obtaining a dust sample based on the dust sample collection mechanism;

[0018] Step S2: the dust image acquisition mechanism acquires an image of the dust sample in step S1 to obtain a dust sample image;

[0019] Step S3: the host computer receives the dust sample image transmitted by the dust image acquisition mechanism, and the data processing module in the host computer performs data processing on the dust sample image.

[0020] Preferably, obtaining the dust sample comprises the following steps:

[0021] Step S101: Start the air pump to add high-pressure gas to the gas storage tank;

[0022] Step S102: starting the air intake fan, so that the dust-laden airflow in the coal mine enters the ventilation pipe under the action of the air intake fan and flows in the ventilation pipe;

[0023] Step S103: The air intake fan is turned off, and the dust-laden airflow begins to settle under the action of its own weight, and some dust falls onto the glass slide;

[0024] Step S104: activating the electromagnet, the slide holder drives the cover glass to move downward, the cover glass and the slide come into contact, and a dust sample is obtained;

[0025] Obtaining the dust sample image comprises the following steps:

[0026] Step S201: Start the light source, and illuminate the slide obliquely upward;

[0027] Step S202: starting an industrial camera to capture images of dust particles on the dust sample to obtain a dust sample image;

[0028] The data processing module performs data processing on the dust sample image, including the following steps:

[0029] Step S301: The industrial camera transmits the dust sample image to the transmission module of the host computer;

[0030] Step S302: the transmission module of the host computer transmits the received dust sample image to the image processing module of the host computer;

[0031] Step S303: The image processing module of the host computer performs image processing on the received dust sample image and obtains dust concentration and dust particle size distribution data;

[0032] Step S304: the image processing module of the host computer transmits the dust concentration and dust particle size distribution data to the display module of the host computer for display.

[0033] Compared with the prior art, the present invention has the following advantages:

[0034] 1. The present invention includes an industrial camera and a host computer. The industrial camera collects dust sample images and transmits the collected sample images to the image processing module of the host computer for processing. The image processing module in the present invention is established after model training of the initial image processing module. The image processing module can synchronously analyze the dust concentration and dust particle size distribution data based on the dust sample image. The staff can have a more comprehensive understanding of the dust situation in the underground working environment of the coal mine based on the dust concentration and dust particle size distribution data, and take timely measures to control it, thereby improving the health protection of the underground coal mine workers.

[0035] 2. The present invention sets the detection device in a box and provides a handle on the box to facilitate the staff to move the device to different detection locations for use. The box can also be connected to a slide rail for continuous measurement along the way. The compact design of the detection device improves the applicability of the detection device.

[0036] 3. The host computer of the present invention is also provided with a plurality of buttons, which can issue instructions to the dust sampling mechanism, the dust image acquisition mechanism and the dust discharge mechanism. The host computer controls the various components in the device to perform corresponding actions, thereby improving the convenience and work efficiency of the detection device operation.

[0037] The present invention is further described in detail below through the accompanying drawings and examples. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 Schematic diagram of the structure of the detection device of the present invention;

[0039] Figure 2 This is a schematic structural diagram of the detection device of the present invention without the housing;

[0040] Figure 3 This is a schematic diagram of the external structure of the box body of the present invention;

[0041] Figure 4 It is a schematic diagram of the bottom structure of the ventilation pipe of the present invention.

[0042] Description of the accompanying drawings:

[0043] 1—base; 2—manual valve; 3—exhaust hose;

[0044] 4—Camera bracket; 5—Solenoid valve; 6—Light source bracket;

[0045] 7—Nozzle cover; 8—Intake fan; 9—Ventilation pipe;

[0046] 10—cover glass; 11—electromagnet; 12—slide rack;

[0047] 13—Ventilation pipe fixing bracket; 14—Industrial camera; 15—Light source;

[0048] 16—electromagnet fixing bracket; 17—slide rail; 18—punching plate;

[0049] 19—Pressure gauge; 20—Air pump; 21—Inlet hose;

[0050] 22—gas storage tank; 23—second tank door; 24—first tank door;

[0051] 25—Box body; 26—First connecting ring; 27—Handle;

[0052] 28—high-pressure nozzle; 29—flip baffle; 30—glass slide;

[0053] 31—Ventilation port; 32—First mounting port; 33—Second mounting port;

[0054] 34—housing frame; 35—battery; 36—host computer;

[0055] 37—Second connecting ring; 38—Air outlet; 39—Button. DETAILED DESCRIPTION

[0056] like Figures 1 to 3 As shown, the present invention includes a box body 25, a punching plate 18 is fixed in the middle of the box body 25, and the punching plate 18 divides the inner cavity of the box body 25 into an upper cavity and a lower cavity. A dust sample collection mechanism for collecting dust samples is provided in the upper cavity of the box body 25, and a dust image collection mechanism for collecting dust samples to obtain dust sample images is provided on the lower side of the dust sample collection mechanism. A host computer 36 is provided on the side wall of the upper cavity of the box body 25. The host computer 36 includes a transmission module, an image processing module and a display module. The transmission module is communicatively connected to the image processing module. The transmission module is used to send and receive data. After receiving the dust sample image transmitted by the dust image acquisition mechanism, the transmission module transmits the dust sample image to the image processing module. The image processing module performs image analysis on the received dust sample image to obtain dust concentration and dust particle size distribution data. The output end of the image processing module is connected to the input end of the display module. The image processing module transmits the dust concentration and dust particle size distribution data to the display module for display. A dust discharge mechanism for discharging dust in the dust sample acquisition mechanism is provided in the lower cavity of the box body 25.

[0057] The punching plate 18 is fixedly connected to the inner wall of the box body 25 so as to support the dust sample collection mechanism and the dust image collection mechanism located on the upper side of the punching plate 18. When performing dust particle size and concentration detection, the dust sample is first obtained by the dust sample collection mechanism, and then the dust image collection mechanism collects the dust sample and obtains the dust sample image. The dust image collection mechanism transmits the dust sample image to the transmission module of the host computer 36, and the transmission module of the host computer 36 transmits the received dust sample image to the image processing module of the host computer 36. The image processing module of the host computer 36 processes the dust sample image. After image analysis and processing, the dust concentration and dust particle size distribution data are obtained. The image processing module transmits the dust concentration and dust particle size distribution data to the display module for display. The staff can understand the dust situation under the mine through the dust particle size and concentration data displayed on the display module; after the data processing is completed, the dust discharge mechanism is started, and the dust discharge mechanism discharges high-pressure gas to discharge the dust in the dust sample collection mechanism to the outside of the dust sample collection mechanism, so that the dust sample collection mechanism remains clean. A handle 27 is also provided on the outer top of the box body 25, and the handle 27 can be used to conveniently carry the box body 25 to any point to be detected.

[0058] like Figure 3As shown, the host computer 36 further includes a plurality of buttons 39 , and the buttons 39 on the host computer 36 can be operated to issue instructions to control the start or stop of the dust sample collection mechanism, the dust image collection mechanism and the dust discharge mechanism.

[0059] The multiple buttons 39 include a power button and multiple setting buttons. The power button can be used to control the power on and off of the device. The multiple setting buttons can be used to set the collection mode of the dust sample collection mechanism, the image collection method of the dust image collection mechanism, the image magnification, the number of collected images, etc., and the working time of the dust discharge mechanism can also be set.

[0060] The image processing module includes a preprocessing model, a feature extraction model, and a neural network model. The neural network model is trained according to the following steps:

[0061] Step S1: Build an experimental platform and obtain a dust sample image set and label dataset based on the experimental platform;

[0062] The experimental platform built is a dust image acquisition experimental platform, which includes a dust particle image detection device, a blower, a screw feeder, a Venturi gas-solid mixer, a total dust concentration sensor, a respirable dust concentration sensor, etc. Through this experimental platform, a large number of particle sample images and label data sets with different total dust and respirable dust concentrations are obtained.

[0063] The dust sample image set is the dust particle images collected through the experimental platform, and the label data set is the dust particle size mass ratio and concentration data corresponding to the collected dust particle images. The neural network model integrates deep learning and multi-task learning. It takes dust particle images as input and simultaneously outputs the mass concentration of total dust, respirable dust and the mass ratio of coarse dust, fine dust and micro dust, realizing the simultaneous rapid detection of dust concentration and mass dispersion (particle size), which is convenient for revealing the sedimentation and diffusion laws of dust, reducing the corresponding speed of sedimentation and dust removal facilities, preventing pneumoconiosis, and protecting workers' health, providing complete and reliable data support for precise dust control.

[0064] The neural network model introduces the idea of ​​multi-task learning to facilitate the multi-task prediction requirements of total dust mass concentration, respirable dust mass concentration and the mass proportion of coarse dust, fine dust and micro dust. Considering that respirable dust is essentially a type of small particle dust, the respirable dust mass concentration is converted into its mass proportion and trained together with the mass proportion of coarse dust, fine dust and micro dust as a task branch. Therefore, the neural network model has two task branches: Task 1 is the prediction of total dust concentration, and Task 2 is the prediction of the mass proportion of different dust particle sizes.

[0065] The two task branches do not exist completely independently in the neural network model. The two main layers of the neural network model are the shared layer and the Set Transformer layer. The shared layer optimizes information sharing between tasks, thereby solving the representation consistency problem in multi-task learning; the Set Transformer layer improves the ability to handle dynamic changes in the number of particles through the attention pooling module; the Set Transformer converts the variable-length geometric feature input into a fixed-length high-dimensional feature representation, and then inputs it into the second task branch after feature splicing with the shared layer output; at the same time, the shared layer output is directly input into the first task branch. The two task branches are fully connected neural networks.

[0066] Step S2: preprocessing the dust sample image set using the preprocessing model, extracting features from the preprocessed image using the feature extraction model, and performing correlation analysis and redundancy elimination on the feature parameters and the total dust concentration to obtain a feature set for input into the neural network model;

[0067] The preprocessing model performs image denoising, image enhancement, image segmentation and other preprocessing on the dust particle image to reduce the impact of vibration, lighting, electromagnetic interference and other problems on the quality of the dust particle image. Then, the feature extraction model is used to extract features from the preprocessed image as the input of the neural network model. The input of the neural network model is the features extracted from the image. Before the neural network model, there is an initial image preprocessing model and feature extraction model for image preprocessing and feature extraction.

[0068] Step S3: The neural network model is trained according to the feature set and label data set of the input neural network model to obtain a trained neural network model.

[0069] The neural network model training used the Adam solver and the Optuna hyperparameter tuning algorithm, and a trained neural network model was obtained after a large number of iterative optimizations. The trained neural network model was deployed to the host computer as the image processing module of the host computer, and an experimental platform for dust concentration and mass distribution image detection was built. Six detection points were set along the dust diffusion direction on the return air side, and images were collected at each measurement point for preprocessing. The neural network model was applied to obtain the total dust, respirable dust concentration, and the mass proportion of coarse dust, fine dust, and micro dust (particle size distribution). Each measurement point was tested 5 times and the average value was taken as the predicted value. At the same time, the dry laser particle size was used to measure the dust concentration and mass distribution of the dust. The analyzer and the direct-reading dust concentration meter were used for synchronous detection as the measurement values ​​of the standard instruments, and the average relative error was obtained by comparing the predicted values ​​with the measurement values ​​of the standard instrument. The device was compared with existing instruments: compared with the direct-reading dust concentration meter, the average relative error of the total dust concentration prediction was 7.3001%, and the average relative error of the respirable dust concentration prediction was 7.3201%; compared with the dry laser particle size analyzer, the average relative error of the particle size distribution was 7.3137%; after processing with the trained neural network model, the error of the dust particle size was kept within 6.63%, and the error of the dust concentration was kept within 12.37%.

[0070] like Figures 1 to 3 As shown, a dust channel is also provided on the upper part of the upper cavity of the box body 25, and the dust channel includes an air intake fan 8, a ventilation pipe 9, a ventilation port 31 and an air outlet 38. The ventilation port 31 and the air outlet 38 are respectively arranged on the left and right sides of the box body 25, and the air intake fan 8 is installed on the ventilation port 31, and the ventilation pipe 9 is installed between the ventilation port 31 and the air outlet 38, and the left and right end portions of the ventilation pipe 9 are respectively connected to the ventilation port 31 and the air outlet 38, and the upper and lower sides of the ventilation pipe 9 are respectively provided with a first mounting port 32 and a second mounting port 33, and the first mounting port 32 is located on the upper side of the second mounting port 33, and the lower side of the ventilation pipe 9 is fixedly connected to a ventilation pipe fixing frame 13, and the ventilation pipe fixing frame 13 is fixedly connected to the punching plate 18 at one end away from the ventilation pipe 9.

[0071] The ventilation pipe 9 is a square tube and is white in color. The size of the pipe openings on the left and right ends of the ventilation pipe 9 matches the size of the ventilation opening 31 and the air outlet 38. The left and right ends of the ventilation pipe 9 are respectively connected to the ventilation opening 31 and the air outlet 38 on the left and right sides of the box body 25 to form a dust passage. The first mounting opening 32 and the second mounting opening 33 are both located in the middle of the ventilation pipe 9. There are two ventilation pipe fixing frames 13 and they are an eight-shaped structure. The two ventilation pipe fixing frames 13 are respectively located at the left and right ends of the punching plate 18. The bottoms of the two ventilation pipe fixing frames 13 are fixedly connected to the punching plate 18. The tops of the two ventilation pipe fixing frames 13 are respectively located on the left and right sides of the second mounting opening 33. The top of each ventilation pipe fixing frame 13 is fixedly connected to the bottom of the ventilation pipe 9.

[0072] There are two ventilation pipe fixing frames 13 and they are in the shape of a letter "Z". The upper part of the ventilation pipe fixing frame 13 is a fastening frame, which is sleeved on the outside of the ventilation pipe 9. The ventilation pipe fixing frame 13 is fastened to the ventilation pipe 9 through the fastening frame. The lower part of the ventilation pipe fixing frame 13 is a bracket, the bottom of the bracket is fixedly connected to the punching plate 18, and the top of the bracket is fixedly connected to the bottom of the fastening frame and supports the fastening frame and the ventilation pipe 9 upward; when the ventilation pipe 9 is a circular structure, the left and right side ends of the ventilation pipe 9 are respectively connected to the ventilation port 31 and the air outlet 38. The inner diameter of the fastening frame of the letter "Z" shaped ventilation pipe fixing frame 13 is consistent with the outer diameter of the ventilation pipe 9 and the fastening frame is sleeved on the outside of the ventilation pipe 9.

[0073] like Figure 2 、 Figure 4 As shown, the dust sample collection mechanism includes a slide rack 12, a slide rail 17, a slide glass 30 and a cover glass 10. There are two slide rails 17, and the two slide rails 17 are fixedly arranged on the front and rear sides of the top of the punching plate 18 respectively. Each slide rail 17 is provided with an electromagnet fixing frame 16 that slides along the slide rail 17. Each electromagnet fixing frame 16 is installed with an electromagnet 11. The two electromagnets 11 are respectively located on the front and rear sides of the ventilation pipe 9. The slide rack 12 is a T-shaped structure. The vertical end portion of the slide rack 12 is connected to the cover glass 10 so that the cover glass 10 is located in the first mounting port 32. The horizontal end portion of the slide rack 12 is respectively provided with a first sleeve ring 26 and a second sleeve ring 37. The first sleeve ring 26 and the second sleeve ring 37 are respectively sleeved on the electromagnets 11 located on the front and rear sides of the ventilation pipe 9. When the electromagnet 11 is energized, it drives the slide rack 12 to move along the electromagnet 11, and the slide glass 30 is fixed in the second mounting port 33.

[0074] The size of the glass slide 30 is 40mm*40mm, which matches the size of the second mounting opening 33 and is fixedly installed in the second mounting opening 33; the electromagnet fixing frame 16 slides left and right along the slide rail 17, driving the glass slide rack 12 on the electromagnet 11 to move left and right to adjust the position of the cover glass 10. When the cover glass 10 on the glass slide rack 12 is located in the first mounting opening 32, the sliding of the electromagnet fixing frame 16 is stopped, and the electromagnet fixing frame 16 is locked to prevent the cover glass 10 and the glass slide 30 from being misaligned due to the movement of the electromagnet fixing frame 16; after the dust in the ventilation pipe 9 is settled, the electromagnet 11 is energized, and the glass slide rack 12 drives the cover glass 10 to move downward until the cover glass 10 contacts the glass slide 30 to obtain a dust glass slide, at which time the glass slide rack 12 stops moving.

[0075] like Figure 2 、 Figure 4As shown, the dust image acquisition mechanism includes a light source 15, a light source bracket 6, an industrial camera 14 and a camera bracket 4. The light source bracket 6 is a hollow bracket and the top and bottom of the bracket are open. The light source 15 is a strip light source and is installed at the bottom opening of the light source bracket 6. The top of the light source bracket 6 is installed at the second installation port 33 and is fixedly connected to the pipe wall of the ventilation pipe 9. The industrial camera 14 is located in the light source bracket 6 and the industrial camera 14 is installed on the camera bracket 4. The camera bracket 4 is a square support frame. The bottom of the square support frame is fixedly connected to the punching plate 18. The industrial camera 14 is installed on the top of the square support frame and the industrial camera 14 is 50 mm away from the slide 30.

[0076] The top opening of the light source bracket 6 matches the size of the second mounting port 33, and the top of the light source bracket 6 is fixedly connected to the pipe wall of the ventilation pipe 9 located at the second mounting port 33, so that the second mounting port 33 is located inside the light source bracket 6; the left and right end portions of the bottom of the light source bracket 6 are inclined toward the inner side of the light source bracket 6 to form an upwardly inclined inwardly inclined surface, the size of which is consistent with the size of the light source 15. There are two light sources 15, and the two light sources 15 are respectively installed on the two inwardly inclined surfaces at the bottom of the light source bracket 6, so that the light source 15 illuminates the glass slide obliquely upward; the industrial camera 14 is located on the upper side of the light source 15, so that the light source 15 can illuminate the front side of the glass slide; when collecting dust samples on the glass slide, the light source 15 is powered on to illuminate the glass slide, the microscope lens of the industrial camera 14 amplifies the dust particles on the dust sample, and collects the dust particle image through the CCD sensor connected to the microscope lens to obtain a dust sample image. The industrial camera 14 transmits the dust sample image to the transmission module of the host computer 36, and the light source 15 is powered off.

[0077] like Figure 1 、 Figure 2 and Figure 4 As shown, the dust discharge mechanism includes a pressure gauge 19, an air pump 20, an air tank 22 and a nozzle cover 7. The pressure gauge 19 is installed on the upper side of the air pump 20 to detect the gas pressure in the air tank 22. The air pump 20 is installed on the upper side of the air tank 22. An air intake hose 21 is also installed on the air pump 20. The end of the air intake hose 21 away from the air pump 20 is connected to the air tank 22. The air pump 20 delivers gas to the air tank 22 through the air intake hose 21. An exhaust hose is also installed on the top of the air tank 22. 3. A manual valve 2 is provided at the end of the exhaust hose 3 near the gas tank 22. The nozzle cover 7 is located at the bottom of the ventilation pipe 9 and is fixedly connected to the ventilation pipe 9 by bolts. A high-pressure nozzle 28 and a flip baffle 29 are installed in the nozzle cover 7. The flip baffle 29 is installed on the upper side of the high-pressure nozzle 28. The left end of the flip baffle 29 is rotatably connected to the inner side wall of the nozzle cover 7. A solenoid valve 5 is installed on the lower side of the high-pressure nozzle 28. The solenoid valve 5 is connected to the end of the exhaust hose 3 away from the gas tank 22.

[0078] The gas tank 22 is installed on the base 1 at the bottom of the lower cavity of the box body 25; the pressure gauge 19 detects the gas pressure in the gas tank 22 before and after the equipment performs dust detection. If the gas pressure in the gas tank 22 is lower than the preset value, the air pump 20 is started, and the air pump 20 replenishes gas into the gas tank 22 through the air inlet hose 21; the preset value can be the minimum air pressure of the gas tank 22 set in advance.

[0079] The flip baffle 29 is rectangular and is connected to the left inner wall of the nozzle cover 7 near the ventilation pipe 9 by rotating along its width direction. The lower side of the flip baffle 29 contacts the high-pressure nozzle 28, and the right end of the flip baffle 29 is supported upward by the high-pressure nozzle 28 to contact the bottom of the ventilation pipe 9 to prevent the dust in the ventilation pipe 9 from settling to the outside of the ventilation pipe 9; when it is necessary to discharge the dust particles in the ventilation pipe 9, the manual valve 2 and the solenoid valve 5 are opened, and the high-pressure gas in the gas tank 22 is ejected by the high-pressure nozzle 28 after passing through the exhaust hose 3. The high-pressure gas ejected by the high-pressure nozzle 28 pushes the flip baffle 29 along the nozzle cover The left inner wall of the ventilation pipe 7 rotates, and the right end of the flip baffle 29 rotates upward under the push of the high-pressure gas until it contacts the inner wall of the ventilation pipe 9. At this time, the front and rear ends of the flip baffle 29 are in contact with the inner wall of the ventilation pipe 9, so that the high-pressure gas can only flow along the ventilation pipe 9 to the air outlet 38. Under the action of the high-pressure gas, the dust in the ventilation pipe 9 is discharged from the air outlet 38; when the ventilation pipe 9 is cleaned, the solenoid valve 5 is closed, the manual valve 2 is closed, the high-pressure nozzle 28 stops spraying high-pressure gas, and the flip baffle 29 rotates downward along the left inner wall of the nozzle cover 7 until it is in the same plane with the lower surface of the inner wall of the ventilation pipe 9.

[0080] like Figure 1 、 Figure 3 As shown, a receiving frame 34 is fixedly connected to the inner wall of the lower cavity of the box body 25, a battery 35 is arranged in the receiving frame 34, and a cover plate is provided on the outer wall of the receiving frame 34, which are detachably connected to the box body 25 and the receiving frame 34 respectively, and the battery 35 is electrically connected to the air intake fan 8, the air pump 20, the electromagnet 11, the light source 15, the solenoid valve 5 and the host computer 36 respectively. A first box door 24 and a second box door 23 are also provided on the front side of the box body 25, and the first box door 24 is located on the upper side of the second box door 23.

[0081] The battery 35 is a rechargeable battery. When charging, the cover can be opened and an external power supply can be connected to the charging port of the battery 35 for charging. The battery 35 supplies power to the air intake fan 8, air pump 20, electromagnet 11, light source 15, solenoid valve 5, and host computer 36 respectively.

[0082] A method for detecting dust particle size and concentration based on image analysis, comprising the following steps:

[0083] Step S1: obtaining a dust sample based on the dust sample collection mechanism;

[0084] Obtaining a dust sample involves the following steps:

[0085] Step S101: Start the air pump 20 to store high-pressure gas into the gas storage tank 22;

[0086] Step S102: starting the air intake fan 8, so that the dust-laden airflow in the coal mine enters the ventilation pipe 9 under the action of the air intake fan 8 and flows in the ventilation pipe 9;

[0087] Step S103: The air intake fan 8 is turned off, and the dust-laden airflow begins to settle under the action of its own weight, and some dust falls onto the glass slide 30;

[0088] Step S104: activating the electromagnet 11, the slide holder 12 drives the cover glass 10 to move downward, the cover glass 10 and the slide glass 30 come into contact, and a dust sample is obtained;

[0089] Step S2: the dust image acquisition mechanism acquires an image of the dust sample in step S1 to obtain a dust sample image;

[0090] Obtaining a dust sample image includes the following steps:

[0091] Step S201: Start the light source 15, and illuminate the glass slide 30 obliquely upward;

[0092] Step S202: starting the industrial camera 14 to capture images of dust particles on the dust sample to obtain an image of the dust sample;

[0093] Step S3: the host computer 36 receives the dust sample image transmitted by the dust image acquisition mechanism, and the data processing module in the host computer 36 performs data processing on the dust sample image.

[0094] The data processing module processes the dust sample image data including the following steps:

[0095] Step S301: the industrial camera 14 transmits the dust sample image to the transmission module of the host computer 36;

[0096] Step S302: the transmission module of the host computer 36 transmits the received dust sample image to the image processing module of the host computer 36;

[0097] Step S303: the image processing module of the host computer 36 performs image processing on the received dust sample image and obtains dust concentration and dust particle size distribution data;

[0098] Step S304: the image processing module of the host computer 36 transmits the dust concentration and dust particle size distribution data to the display module of the host computer 36 for display.

[0099] After the dust sample image processing is completed, the dust sample collection mechanism needs to be cleaned. First, the electromagnet 11 is powered off, and the slide rack 12 drives the cover glass 10 to move upward along the electromagnet 11 to the initial position; then the solenoid valve 5 is opened, and the high-pressure gas in the gas tank 22 is ejected by the high-pressure nozzle 28, pushing the flip baffle 29 to open upward and the edge of the flip baffle 29 contacts the inner wall of the ventilation pipe 9. The high-pressure gas flows along the ventilation pipe 9 to the air outlet, and the dust-containing gas in the ventilation pipe 9 is discharged; then the solenoid valve 5 is closed, the high-pressure nozzle 28 stops ejecting high-pressure gas, and the flip baffle 29 falls back to the initial position; finally, the gas pressure in the gas tank 22 is detected by the pressure gauge 19. If the air pressure in the gas tank 22 is lower than the preset value, the air pump 20 is started to transport gas into the gas tank 22.

[0100] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any way. Any simple modification, change and equivalent structural transformation made to the above embodiment based on the technical essence of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. A dust particle size and concentration detection device based on image analysis, characterized by: The invention comprises a box body (25), a punching plate (18) is fixed in the middle of the box body (25), the punching plate (18) divides the inner cavity of the box body (25) into an upper cavity and a lower cavity, a dust sample collection mechanism for collecting dust samples is provided in the upper cavity of the box body (25), a dust image collection mechanism for collecting dust samples to obtain dust sample images is provided on the lower side of the dust sample collection mechanism, a host computer (36) is provided on the side wall of the upper cavity of the box body (25), the host computer (36) comprises a transmission module, an image processing module and a display module, the transmission module and the image processing module are connected to each other. The processing module is communicatively connected and the transmission module is used to send and receive data. The transmission module receives the dust sample image transmitted by the dust image acquisition mechanism and transmits the dust sample image to the image processing module. The image processing module performs image analysis on the received dust sample image to obtain dust concentration and dust particle size distribution data. The output end of the image processing module is connected to the input end of the display module. The image processing module transmits the dust concentration and dust particle size distribution data to the display module for display. A dust discharge mechanism for discharging dust in the dust sample acquisition mechanism is provided in the lower cavity of the box body (25).

2. The dust particle size and concentration detection device based on image analysis according to claim 1, characterized in that: The host computer (36) further comprises a plurality of buttons (39), and the buttons (39) on the host computer (36) can be operated to issue instructions to control the dust sample collection mechanism, the dust image collection mechanism and the dust discharge mechanism to start or stop.

3. The dust particle size and concentration detection device based on image analysis according to claim 1, characterized in that: The image processing module includes a preprocessing model, a feature extraction model and a neural network model, and the neural network model is trained according to the following steps: Step S1: Build an experimental platform and obtain a dust sample image set and label dataset based on the experimental platform; Step S2: preprocessing the dust sample image set using the preprocessing model, extracting features from the preprocessed image using the feature extraction model, and performing correlation analysis and redundancy elimination on the feature parameters and the total dust concentration to obtain a feature set for input into the neural network model; Step S3: The neural network model is trained according to the feature set and label data set of the input neural network model to obtain a trained neural network model.

4. The dust particle size and concentration detection device based on image analysis according to claim 1, characterized in that: The upper part of the upper cavity of the box body (25) is also provided with a dust passage, and the dust passage comprises an air intake fan (8), a ventilation pipe (9), a ventilation port (31) and an air outlet (38), wherein the ventilation port (31) and the air outlet (38) are respectively provided on the left and right sides of the box body (25), the air intake fan (8) is installed on the ventilation port (31), the ventilation pipe (9) is installed between the ventilation port (31) and the air outlet (38), and the left and right end portions of the ventilation pipe (9) are respectively connected to the ventilation port (31) and the air outlet (38), and the upper and lower sides of the ventilation pipe (9) are respectively provided with a first mounting port (32) and a second mounting port (33), wherein the first mounting port (32) is located above the second mounting port (33), and the lower side of the ventilation pipe (9) is fixedly connected to a ventilation pipe fixing frame (13), and the end of the ventilation pipe fixing frame (13) away from the ventilation pipe (9) is fixedly connected to the punching plate (18).

5. The dust particle size and concentration detection device based on image analysis according to claim 4, characterized in that: The dust sample collection mechanism comprises a glass slide rack (12), a slide rail (17), a glass slide (30) and a cover glass (10), wherein the number of the slide rails (17) is two, and the two slide rails (17) are respectively fixedly arranged on the front and rear sides of the top of the punching plate (18), and each of the slide rails (17) is provided with an electromagnet fixing frame (16) sliding along the slide rail (17), and each of the electromagnet fixing frames (16) is installed with an electromagnet (11), and the two electromagnets (11) are respectively located on the front and rear sides of the ventilation pipe (9), and the glass slide rack (12) is a T-shaped structure. The invention relates to a structure in which the vertical end portion of the glass slide rack (12) is connected to the cover glass (10) so that the cover glass (10) is located in the first installation opening (32); the horizontal end portion of the glass slide rack (12) is respectively provided with a first sleeve ring (26) and a second sleeve ring (37); the first sleeve ring (26) and the second sleeve ring (37) are respectively sleeved on the electromagnets (11) located on the front and rear sides of the ventilation pipe (9); the electromagnet (11) drives the glass slide rack (12) to move along the electromagnet (11) when energized, and the glass slide (30) is fixedly arranged in the second installation opening (33).

6. The dust particle size and concentration detection device based on image analysis according to claim 4, characterized in that: The dust image acquisition mechanism comprises a light source (15), a light source bracket (6), an industrial camera (14) and a camera bracket (4); the light source bracket (6) is a hollow bracket with both the top and bottom of the bracket open; the light source (15) is a strip light source and is mounted at the bottom opening of the light source bracket (6); the top of the light source bracket (6) is mounted at the second mounting opening (33) and is fixedly connected to the wall of the ventilation pipe (9); the industrial camera (14) is located in the light source bracket (6) and the industrial camera (14) is mounted on the camera bracket (4); the camera bracket (4) is a square support bracket, the bottom of the square support bracket is fixedly connected to the punching plate (18); the industrial camera (14) is mounted on the top of the square support bracket, and the distance between the industrial camera (14) and the slide glass (30) is 50 mm.

7. The dust particle size and concentration detection device based on image analysis according to claim 4, characterized in that: The dust discharge mechanism comprises a pressure gauge (19), an air pump (20), an air storage tank (22) and a nozzle cover (7). The pressure gauge (19) is installed on the upper side of the air pump (20) for detecting the gas pressure in the air storage tank (22). The air pump (20) is installed on the upper side of the air storage tank (22). An air intake hose (21) is also installed on the air pump (20). One end of the air intake hose (21) away from the air pump (20) is connected to the air storage tank (22). The air pump (20) delivers gas to the air storage tank (22) through the air intake hose (21). An exhaust hose (3) is also installed on the top of the air storage tank (22). A manual valve (2) is provided at the end of the exhaust hose (3) close to the gas storage tank (22), the nozzle cover (7) is located at the bottom of the ventilation pipe (9) and is fixedly connected to the ventilation pipe (9) by bolts, a high-pressure nozzle (28) and a flip baffle (29) are installed in the nozzle cover (7), the flip baffle (29) is installed on the upper side of the high-pressure nozzle (28), the left end of the flip baffle (29) is rotatably connected to the inner side wall of the nozzle cover (7), and a solenoid valve (5) is installed on the lower side of the high-pressure nozzle (28), and the solenoid valve (5) is connected to the end of the exhaust hose (3) away from the gas storage tank (22).

8. A dust particle size and concentration detection device based on image analysis according to any one of claims 1 to 7, characterized in that: A receiving frame (34) is fixedly connected to the inner side wall of the lower cavity of the box body (25), a battery (35) is arranged in the receiving frame (34), and a cover plate is provided on the outer side wall of the receiving frame (34), and the cover plate is detachably connected to the box body (25) and the receiving frame (34), respectively. The battery (35) is electrically connected to the intake fan (8), the air pump (20), the electromagnet (11), the light source (15), the electromagnetic valve (5) and the host computer (36), respectively. The front side of the box body (25) is also provided with a first box door (24) and a second box door (23), and the first box door (24) is located on the upper side of the second box door (23).

9. A detection method for a dust particle size and concentration detection device based on image analysis according to any one of claims 1 to 8, characterized in that: The following steps are involved: Step S1: obtaining a dust sample based on the dust sample collection mechanism; Step S2: the dust image acquisition mechanism acquires an image of the dust sample in step S1 to obtain a dust sample image; Step S3: the host computer (36) receives the dust sample image transmitted by the dust image acquisition mechanism, and the image processing module in the host computer (36) processes the dust sample image.

10. The method for detecting dust particle size and concentration based on image analysis according to claim 9, characterized in that: The obtaining of dust samples comprises the following steps: Step S101: Start the air pump (20) to replenish high-pressure gas into the gas storage tank (22); Step S102: starting the air intake fan (8), so that the dust-laden airflow in the coal mine enters the ventilation pipe (9) under the action of the air intake fan (8) and flows in the ventilation pipe (9); Step S103: The air intake fan (8) is turned off, and the dust-laden airflow begins to settle under the action of its own weight, and part of the dust falls onto the glass slide (30); Step S104: the electromagnet (11) is activated, and the slide holder (12) drives the cover glass (10) to move downward, so that the cover glass (10) and the slide glass (30) come into contact with each other, thereby obtaining a dust sample; Obtaining the dust sample image comprises the following steps: Step S201: Start the light source (15), and the light source (15) illuminates the glass slide (30) obliquely upward; Step S202: starting the industrial camera (14), the industrial camera (14) collects images of dust particles on the dust sample to obtain a dust sample image; The data processing module performs data processing on the dust sample image, including the following steps: Step S301: The industrial camera (14) transmits the dust sample image to the transmission module of the host computer (36); Step S302: the transmission module of the host computer (36) transmits the received dust sample image to the image processing module of the host computer (36); Step S303: the image processing module of the host computer (36) performs image processing on the received dust sample image and obtains dust concentration and dust particle size distribution data; Step S304: the image processing module of the host computer (36) transmits the dust concentration and dust particle size distribution data to the display module of the host computer (36) for display.

Images