Method and device for monitoring position state of plow type belt discharger based on machine vision

Abstract

The invention relates to a plough type belt unloader position state monitoring method and device based on machine vision. The plough type belt unloader position state monitoring method comprises the following steps that firstly, images are collected and marked; 2, training a classification model; step 3, a target detection model; 4, deploying the model; 5, monitoring and controlling in real time; step 6, performing exception handling; and step seven, data storage. Through a deep learning algorithm, the system can accurately identify the upper and lower limit positions of the discharger, ensure that the discharger is lifted and lowered in place, reduce material leakage or accumulation caused by inaccurate positions, monitor the state of the discharger in real time, automatically control the discharger to move to a specified position, reduce manual intervention, and improve the working efficiency. And the operation efficiency and the response speed are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of machine vision, and in particular relates to a method and device for monitoring the position status of a plow-type belt unloader based on machine vision. Background Art

[0002] As the "switch" of the coal bunker inlet, the unloader directly determines whether the coal level in the bunker reaches the expected height, which in turn affects the completion of the entire process. If the unloader is not lifted into place, it is easy to intercept part of the coal flow on the belt to the ground or coal bunker on both sides. Conversely, if the unloader is not lowered into place, part of the coal flow is easy to leak out from the bottom of the unloader, affecting production efficiency and production safety.

[0003] At present, traditional methods for monitoring the position of dischargers often rely on manual inspections or simple sensors. Traditional monitoring methods have many disadvantages, including low efficiency, susceptibility to human factors, slow response, limited data recording and analysis capabilities, high costs, etc., and cannot achieve real-time monitoring. With the rapid development of machine vision technology, its application in the field of industrial automation is becoming more and more extensive. Machine vision technology uses cameras to obtain image information of the target environment in real time, and uses deep learning algorithms to analyze and process the images, thereby realizing the functions of identifying, locating, and measuring the target objects. Compared with traditional manual inspections and sensor monitoring methods, machine vision technology has higher accuracy, faster speed, and stronger adaptability. Therefore, in the production process of coal transportation, it is necessary to use machine vision to detect the position and status of the discharger in real time to improve production efficiency. Summary of the invention

[0004] (I) Purpose of the invention

[0005] In order to overcome the above shortcomings, the purpose of the present invention is to provide a method and device for monitoring the position status of a plow-type belt discharger based on machine vision to solve the above technical problems.

[0006] (II) Technical solution

[0007] To achieve the above objectives, the technical solutions provided by this application are as follows:

[0008] A method for monitoring the position status of a plow-type belt unloader based on machine vision comprises the following steps:

[0009] Step 1: Image acquisition and annotation. The image field of view includes the unloader and the rear belt. The unloader is annotated through the target detection data set, and the collected belt image is annotated by category through the classification recognition data set.

[0010] Step 2: training the classification model. Based on the classification identification data set in step 1, the coal-existence data set is divided into a training set and a validation set, and a coal-existence classification model is trained.

[0011] Step 3: Target detection model: Based on the unloader data annotated in the target detection data set in step 1, YoloV10s is used as the target detection model. The YoloV10s model structure mainly includes: a backbone network, a feature enhancement network and a detection head;

[0012] Step 4: Model deployment: deploy the trained model to the AI ​​edge computing server for real-time image processing.

[0013] Step 5: Real-time monitoring and control: The real-time monitored images are transmitted to the deployed model to detect the position of the discharger and the coal flow status. According to the detection results, the discharger is automatically controlled to move to the specified position and a voice reminder is given;

[0014] Step 6: Exception handling: If the discharger does not reach the designated position, the discharger is automatically controlled to the designated position and the evidence image is sent to the platform. If the coal flow is continuously identified, the coal baffle at the bottom of the discharger is determined to be damaged, and the evidence image is sent and the on-site voice reminder is made to replace the coal baffle.

[0015] Step 7: Data storage: storing the image identified as a foreign object in an SD card or hard disk with a larger storage capacity.

[0016] Preferably, in step one, the image is captured by an intrinsically safe camera, and the camera is installed just above the running direction of the belt. The image of the discharger and the belt are captured downward, the discharger occupies 40% of the image area and is centered, the image resolution is not less than 1920*1080, and the captured images are divided into images of the discharger at various heights and images of the normal operation of the belt with and without coal.

[0017] Preferably, the target detection data in step one is annotated using deep learning image labeling software CVAT, the annotation type is a rectangular box, and the annotation category is a discharger.

[0018] Preferably, the classification and recognition data set in step one includes images of empty belts and images of belts with coal. The collected images are cropped to crop the belt area behind the discharger, and the images are separately labeled to mark the categories of belts with coal and belts without coal.

[0019] Preferably, the classification model in step 2 is ResNet18.

[0020] Preferably, the classification model ResNet18 includes the following parts:

[0021] Input layer, receiving the belt image behind the belt unloader;

[0022] Convolution layer, using a 7x7 convolution kernel to perform convolution operation to extract coal image features, with a step size of 2, followed by a maximum pooling layer;

[0023] Residual block, ResNet18 consists of multiple residual blocks, each of which contains three convolutional layers, namely 1x1, 3x3, and 1x1, where 1x1 is used for dimensionality transformation and 3x3 convolution is used to extract features;

[0024] Residual connections, where the input of each residual block is directly added to the output of the block through a short connection, which helps solve the vanishing gradient problem;

[0025] Global average pooling layer, used at the end of the network to reduce the number of parameters of the fully connected layer;

[0026] Fully connected layer, used for coal classification task.

[0027] Preferably, the YoloV10s model specifically includes:

[0028] The backbone network uses a lightweight convolutional neural network to extract features from the input image. It contains multiple convolutional layers and residual connections to improve the efficiency and accuracy of feature extraction.

[0029] The feature enhancement network is located between the backbone network and the detection head and is responsible for fusing feature maps of different scales.

[0030] The detection head uses the idea of ​​decoupling the head to separate the regression branch and the prediction branch, making the network training and reasoning more efficient;

[0031] Loss function, uses a composite loss function, including classification loss, localization loss, and target loss, which work together on the prediction results during training to optimize model performance.

[0032] Preferably, the target detection model is provided with two red lines, namely an upper line and a lower line. The upper line serves as a judgment line for whether the unloader has risen to the right position, and the lower line serves as a judgment line for whether the unloader has descended to the right position. The blue box is the unloader detection result, and the position status of the unloader is judged by the y coordinate of the lower right corner of the unloader detection box.

[0033] Preferably, step five specifically includes:

[0034] Input the image into the target detection model of the unloader to obtain the unloader detection result, and determine the position of the unloader based on the y coordinate value of the lower right corner of the detection frame:

[0035] If the y value of the unloader detection frame is less than the y value of the upper line, it means that the unloader has risen to the position;

[0036] If the y value of the unloader detection box is greater than the y value of the upper line and less than the y value of the lower line, it means that the unloader cannot rise to the right position;

[0037] If the y value of the unloader detection frame is greater than the y value of the lower position line, it means that the unloader has dropped into place;

[0038] On this basis, the position information of the unloader is obtained. If it is detected that the unloader has dropped into place, the belt is checked for the presence or absence of coal according to step two to obtain the classification result of the presence or absence of coal in the current image. If 10 consecutive frames of images are judged to be coal-containing, it means that the coal baffle plate of the unloader is damaged. At this time, the relevant images are uploaded to the system platform and the camera at the corresponding position is linked to give an on-site voice reminder "The unloader is found to be damaged, please deal with it as soon as possible". If it is detected that the unloader cannot be raised into place, the relevant images are uploaded to the system platform and a voice alarm is given to remind the administrator to deal with it as soon as possible. The camera at the corresponding position needs to be linked to give an on-site voice reminder "The unloader is found to be unable to be lifted into place, please deal with it as soon as possible". At this time, the lifting command is sent through the joint control camera to control the unloader to be lifted to the specified position.

[0039] A machine vision-based position status monitoring device for a plow-type belt unloader includes an image acquisition device, a communication transmission device, a data processing device, and a data storage device.

[0040] The image acquisition device is an intrinsically safe camera in the coal mine, which can collect high-definition images of the coal conveyor belt during operation. The surface of the intrinsically safe camera is provided with dust-proof material to prevent the adhesion of dust in the mine and affect the imaging quality of the image. The intrinsically safe camera is provided with a joint-controlled camera. The DO interface of the joint-controlled camera control board is connected to the unloader PLC device to realize voice reminder and control of the unloader. The joint-controlled camera control board can control the opening and closing of the two-way switch;

[0041] Communication transmission equipment includes equipment required for 5G, industrial wifi and other wireless transmission methods, as well as equipment required for USB3.0, Ethernet and other wired transmission methods;

[0042] The edge processing device uses an AI edge computing server, which has the computing power and memory to complete Resnet-based category classification and Yolo-based image object detection tasks;

[0043] The data storage device is an SD card or a hard disk with storage capacity, which is used to store the image identified as a foreign object.

[0044] Beneficial effects:

[0045] It has the following beneficial effects:

[0046] 1. Improve monitoring accuracy and speed: Utilize machine vision technology and deep learning algorithms to achieve high-precision, rapid identification and monitoring of the discharger position and coal flow status, improving monitoring efficiency and accuracy.

[0047] 2. Enhance system adaptability and real-time monitoring capabilities: The system is not affected by ambient light and dust, has strong adaptability, can obtain and process image information in real time, and realize real-time monitoring of the discharger position and coal flow status.

[0048] 3. Reduce costs and improve safety: Compared with traditional monitoring methods, machine vision technology reduces human resource investment and equipment maintenance costs, while improving production efficiency and safety through automatic control and voice reminder functions.

[0049] 4. Intelligent control and exception handling: The monitoring results can automatically control the unloader to move to the specified position, and provide voice reminders and upload evidence images when an abnormality is detected, realizing intelligent exception handling.

[0050] 5. Data recording, analysis and storage: The system can record and analyze monitoring data to provide support for production management and decision-making, and store images identified as foreign objects in large-capacity storage devices for subsequent analysis and processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 It is a work flow chart of the present invention. DETAILED DESCRIPTION

[0052] In order to make the purpose, technical solutions and advantages of the present invention more clear, the following is a detailed description of the present invention in conjunction with the specific implementation methods and with reference to the attached drawings. Figure 1 , the present invention is further described in detail. It should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present invention. In addition, in the following description, descriptions of well-known structures and technologies are omitted to avoid unnecessary confusion of the concept of the present invention.

[0053] The present invention provides a method for monitoring the position state of a plow-type belt unloader based on machine vision, comprising the following steps:

[0054] Step 1: Image acquisition and annotation. The image field of view includes the discharger and the rear belt. In step 1, the image is acquired by an intrinsically safe camera. The camera is installed just above the running direction of the belt, and the images of the discharger and the belt are acquired downward. The discharger occupies 40% of the image area and is centered. The image resolution is not less than 1920*1080. The acquired images are divided into images of the discharger at various heights and images of the belt with and without coal during normal operation. The discharger is annotated through the target detection data set, and the acquired belt images are annotated by category through the classification recognition data set.

[0055] The target detection data described in step 1 is annotated using the deep learning image labeling software CVAT, the annotation type is a rectangular frame, and the annotation category is a discharger;

[0056] The classification and recognition data set includes an empty belt image and an image of a belt with coal. The collected images are cropped to crop the belt area behind the discharger, and the images are labeled to mark the categories of coal and no coal.

[0057] Step 2: training a classification model. Based on the classification recognition data set in step 1, the coal-existence data set is divided into a training set and a validation set, and a coal-existence classification model is trained. The classification model in step 2 is ResNet18.

[0058] The classification model ResNet18 includes the following parts:

[0059] Input layer, receiving the belt image behind the belt unloader;

[0060] Convolution layer, using a 7x7 convolution kernel to perform convolution operation to extract coal image features, with a step size of 2, followed by a maximum pooling layer;

[0061] Residual block, ResNet18 consists of multiple residual blocks, each of which contains three convolutional layers, namely 1x1, 3x3, and 1x1, where 1x1 is used for dimensionality transformation and 3x3 convolution is used to extract features;

[0062] Residual connections, where the input of each residual block is directly added to the output of the block through a short connection, which helps solve the vanishing gradient problem;

[0063] Global average pooling layer, used at the end of the network to reduce the number of parameters of the fully connected layer;

[0064] Fully connected layer, used for coal classification task.

[0065] Step three: target detection model, based on the unloader data annotated in the target detection data set in step one. The target detection model is provided with two red lines, namely the upper line and the lower line. The upper line is used as the judgment line for whether the unloader has risen to the right position, and the lower line is used as the judgment line for whether the unloader has descended to the right position. The blue box is the unloader detection result, and the position status of the unloader is judged by the y coordinate of the lower right corner of the unloader detection box.

[0066] YoloV10s is used as the target detection model. The YoloV10s model structure mainly includes: a backbone network, a feature enhancement network and a detection head;

[0067] The YoloV10s model specifically includes:

[0068] The backbone network uses a lightweight convolutional neural network to extract features from the input image. It contains multiple convolutional layers and residual connections to improve the efficiency and accuracy of feature extraction.

[0069] The feature enhancement network is located between the backbone network and the detection head and is responsible for fusing feature maps of different scales.

[0070] The detection head uses the idea of ​​decoupling the head to separate the regression branch and the prediction branch, making the network training and reasoning more efficient;

[0071] Loss function, uses a composite loss function, including classification loss, localization loss, and target loss, which work together on the prediction results during training to optimize model performance.

[0072] Step 4: Model deployment: deploy the trained model to the AI ​​edge computing server for real-time image processing.

[0073] Step 5: Real-time monitoring and control: The real-time monitored images are transmitted to the deployed model to detect the position of the discharger and the coal flow status. According to the detection results, the discharger is automatically controlled to move to the specified position and a voice reminder is given;

[0074] Step 5 specifically includes:

[0075] Input the image into the target detection model of the unloader to obtain the unloader detection result, and determine the position of the unloader based on the y coordinate value of the lower right corner of the detection frame:

[0076] If the y value of the unloader detection frame is less than the y value of the upper line, it means that the unloader has risen to the position;

[0077] If the y value of the unloader detection box is greater than the y value of the upper line and less than the y value of the lower line, it means that the unloader cannot rise to the right position;

[0078] If the y value of the unloader detection frame is greater than the y value of the lower position line, it means that the unloader has dropped into place;

[0079] On this basis, the position information of the unloader is obtained. If it is detected that the unloader has dropped into place, the belt is checked for the presence or absence of coal according to step two to obtain the classification result of the presence or absence of coal in the current image. If 10 consecutive frames of images are judged to be coal-containing, it means that the coal baffle plate of the unloader is damaged. At this time, the relevant images are uploaded to the system platform and the camera at the corresponding position is linked to give an on-site voice reminder "The unloader is found to be damaged, please deal with it as soon as possible". If it is detected that the unloader cannot be raised into place, the relevant images are uploaded to the system platform and a voice alarm is given to remind the administrator to deal with it as soon as possible. The camera at the corresponding position needs to be linked to give an on-site voice reminder "The unloader is found to be unable to be lifted into place, please deal with it as soon as possible". At this time, the lifting command is sent through the joint control camera to control the unloader to be lifted to the specified position.

[0080] Step 6: Exception handling: If the discharger does not reach the designated position, the discharger is automatically controlled to the designated position and the evidence image is sent to the platform. If the coal flow is continuously identified, the coal baffle at the bottom of the discharger is determined to be damaged, and the evidence image is sent and the on-site voice reminder is made to replace the coal baffle.

[0081] Step 7: Data storage: storing the image identified as a foreign object in an SD card or hard disk with a larger storage capacity.

[0082] A machine vision-based position status monitoring device for a plow-type belt unloader includes an image acquisition device, a communication transmission device, a data processing device, and a data storage device.

[0083] The image acquisition device is an intrinsically safe camera in the coal mine, which can collect high-definition images of the coal conveyor belt during operation. The surface of the intrinsically safe camera is provided with dust-proof material to prevent the adhesion of dust in the mine and affect the imaging quality of the image. The intrinsically safe camera is provided with a joint-controlled camera. The DO interface of the joint-controlled camera control board is connected to the unloader PLC device to realize voice reminder and control of the unloader. The joint-controlled camera control board can control the opening and closing of the two-way switch;

[0084] Communication transmission equipment includes equipment required for 5G, industrial wifi and other wireless transmission methods, as well as equipment required for USB3.0, Ethernet and other wired transmission methods;

[0085] The edge processing device uses an AI edge computing server, which has the computing power and memory to complete Resnet-based category classification and Yolo-based image object detection tasks;

[0086] The data storage device is an SD card or a hard disk with storage capacity, which is used to store the image identified as a foreign object.

[0087] The present invention provides a method and device for monitoring the position status of a plow-type belt discharger based on machine vision. The invention uses an intrinsically safe camera to collect images of the discharger and images of the belt coal flow in advance to establish an image data set. The images in the image data set are used to train a discharger position detection model and a coal-presence classification model. Among them, the discharger detection model includes two types of samples: a single-sided discharger and a double-sided discharger, and the target detection model is trained using all samples. First, the discharger is detected on the image, and whether the discharger is stopped is determined based on the states of the previous and next frames, and whether the discharger is in a specified position is determined based on the detection results. If the discharger is not in the specified position, the discharger is automatically controlled to the specified position, and the evidence image is sent to the platform. If the discharger drops into place, the belt image behind the discharger is cropped, and the classification model is used to classify and identify whether there is coal. When coal is identified for a continuous period of time, it is determined that the coal retaining plate at the bottom of the discharger is damaged, and the evidence image is sent to the platform and the on-site voice reminder is linked to replace the coal retaining plate in time. Through the deep learning algorithm, the system of the present invention can accurately identify the upper and lower limit positions of the unloader, ensure that the unloader is lifted and lowered into place, reduce material leakage or accumulation due to inaccurate positioning, monitor the status of the unloader in real time, and automatically control the unloader to move to the specified position, reducing manual intervention and improving operating efficiency and response speed.

[0088] The present invention utilizes machine vision technology and deep learning algorithms to achieve high-precision, rapid identification and monitoring of the position of the discharger and the state of coal flow, thereby improving the efficiency and accuracy of monitoring. The system is not affected by ambient light and dust, has strong adaptability, and can acquire and process image information in real time to achieve real-time monitoring of the position of the discharger and the state of coal flow. Compared with traditional monitoring methods, machine vision technology reduces human resource investment and equipment maintenance costs, while improving production efficiency and safety through automatic control and voice reminder functions. The monitoring results can automatically control the discharger to move to the specified position, and provide voice reminders and upload evidence images when an abnormality is detected, thus realizing intelligent abnormality handling. The system can record and analyze monitoring data, provide support for production management and decision-making, and store images identified as foreign objects in large-capacity storage devices for subsequent analysis and processing.

[0089] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0090] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for monitoring the position status of a plow-type belt unloader based on machine vision, characterized in that: The following steps are involved: Step 1: Image acquisition and annotation. The image field of view includes the unloader and the rear belt. The unloader is annotated through the target detection data set, and the collected belt image is annotated by category through the classification recognition data set. Step 2: training the classification model. Based on the classification identification data set in step 1, the coal-existence data set is divided into a training set and a validation set, and a coal-existence classification model is trained. Step 3: Target detection model: Based on the unloader data annotated in the target detection data set in step 1, YoloV10s is used as the target detection model. The YoloV10s model structure mainly includes: a backbone network, a feature enhancement network and a detection head; Step 4: Model deployment: deploy the trained model to the AI ​​edge computing server for real-time image processing. Step 5: Real-time monitoring and control: The real-time monitored images are transmitted to the deployed model to detect the position of the discharger and the coal flow status. According to the detection results, the discharger is automatically controlled to move to the specified position and a voice reminder is given; Step 6: Exception handling: If the discharger does not reach the designated position, the discharger is automatically controlled to the designated position and the evidence image is sent to the platform. If the coal flow is continuously identified, the coal baffle at the bottom of the discharger is determined to be damaged, and the evidence image is sent and the on-site voice reminder is made to replace the coal baffle. Step 7: Data storage: storing the image identified as a foreign object in an SD card or hard disk with a larger storage capacity.

2. The method for monitoring the position status of a plow type belt unloader based on machine vision according to claim 1, characterized in that: In the step 1, images are collected by an intrinsically safe camera. The camera is installed just above the running direction of the belt, and the images of the discharger and the belt are collected downward. The discharger occupies 40% of the image area and is centered. The image resolution is not less than 1920*1080. The collected images are divided into images of the discharger at various heights and images of the belt with and without coal during normal operation.

3. The method for monitoring the position status of a plow type belt unloader based on machine vision according to claim 1, characterized in that: The target detection data described in step 1 is annotated using the deep learning image labeling software CVAT, the annotation type is a rectangular box, and the annotation category is a discharger.

4. The method for monitoring the position status of a plow type belt unloader based on machine vision according to claim 1, characterized in that: The classification and recognition data set described in step one includes images of empty belts and images of belts with coal. The collected images are cropped to crop out the belt area behind the discharger, and the images are labeled separately to mark the categories of belts with coal and belts without coal.

5. The method for monitoring the position status of a plow type belt unloader based on machine vision according to claim 1, characterized in that: The classification model in step 2 is ResNet18.

6. The method for monitoring the position status of a plow type belt unloader based on machine vision according to claim 5, characterized in that: The classification model ResNet18 includes the following parts: an input layer, receiving the belt image behind the belt unloader; Convolution layer, using a 7x7 convolution kernel to perform convolution operation to extract coal image features, with a step size of 2, followed by a maximum pooling layer; Residual block, ResNet18 consists of multiple residual blocks, each of which contains three convolutional layers, namely 1x1, 3x3, and 1x1, where 1x1 is used for dimensionality transformation and 3x3 convolution is used to extract features; Residual connections, where the input of each residual block is directly added to the output of the block through a short connection, which helps solve the vanishing gradient problem; Global average pooling layer, used at the end of the network to reduce the number of parameters of the fully connected layer; Fully connected layer, used for coal classification task.

7. The method for monitoring the position status of a plow type belt unloader based on machine vision according to claim 1, characterized in that: The YoloV10s model specifically includes: The backbone network uses a lightweight convolutional neural network to extract features from the input image. It contains multiple convolutional layers and residual connections to improve the efficiency and accuracy of feature extraction. The feature enhancement network is located between the backbone network and the detection head and is responsible for fusing feature maps of different scales. The detection head uses the idea of ​​decoupling head to separate the regression branch and the prediction branch, making the network training and reasoning more efficient; Loss function, uses a composite loss function, including classification loss, localization loss, and target loss, which work together on the prediction results during training to optimize model performance.

8. The method for monitoring the position status of a plow type belt unloader based on machine vision according to claim 1, characterized in that: The target detection model is provided with two red lines, namely the upper line and the lower line. The upper line is used as the judgment line for whether the unloader has risen to the right position, and the lower line is used as the judgment line for whether the unloader has descended to the right position. The blue box is the unloader detection result, and the position status of the unloader is judged by the y coordinate of the lower right corner of the unloader detection box.

9. The method for monitoring the position status of a plow type belt unloader based on machine vision according to claim 8, characterized in that: Step 5 specifically includes: Input the image into the target detection model of the unloader to obtain the unloader detection result, and determine the position of the unloader based on the y coordinate value of the lower right corner of the detection frame: If the y value of the unloader detection frame is less than the y value of the upper line, it means that the unloader has risen to the position; If the y value of the unloader detection box is greater than the y value of the upper line and less than the y value of the lower line, it means that the unloader cannot rise to the right position; If the y value of the unloader detection frame is greater than the y value of the lower position line, it means that the unloader has dropped into place; On this basis, the position information of the unloader is obtained. If it is detected that the unloader has dropped into place, the belt is detected for coal or not according to step 2 to obtain the coal or not classification result of the current image. If 10 consecutive frames of images are judged to be coal, it means that the coal baffle of the unloader is damaged. At this time, the relevant images are uploaded to the system platform and the camera at the corresponding position is linked to give an on-site voice reminder "The unloader is found to be damaged, please deal with it as soon as possible". If it is detected that the unloader cannot be raised into place, the relevant images are uploaded to the system platform and a voice alarm is given to remind the administrator to deal with it as soon as possible. It is necessary to link the camera at the corresponding position to give an on-site voice reminder "The unloader is found to be not lifted into place, please deal with it as soon as possible". At this time, the lifting command is sent through the joint control camera to control the unloader to be lifted to the specified position.

10. A machine vision-based position status monitoring device for a plow-type belt unloader, characterized in that: Including image acquisition equipment, communication transmission equipment, data processing equipment, data storage equipment, The image acquisition device is an intrinsically safe camera in the coal mine, which can collect high-definition images of the coal conveyor belt during operation. The surface of the intrinsically safe camera is provided with dust-proof material to prevent the adhesion of dust in the mine and affect the imaging quality of the image. The intrinsically safe camera is provided with a joint-controlled camera. The DO interface of the joint-controlled camera control board is connected to the unloader PLC device to realize voice reminder and control of the unloader. The joint-controlled camera control board can control the opening and closing of the two-way switch; Communication transmission equipment includes equipment required for 5G, industrial wifi and other wireless transmission methods, as well as equipment required for USB3.0, Ethernet and other wired transmission methods; The edge processing device uses an AI edge computing server, which has the computing power and memory to complete Resnet-based category classification and Yolo-based image object detection tasks; The data storage device is an SD card or a hard disk with storage capacity, which is used to store the image identified as a foreign object.

Images