Method for identifying coal flow blocking state of scraper conveyor on fully mechanized coal mining face of coal mine

Abstract

The invention provides a coal mine fully mechanized coal mining face scraper machine coal flow blocking state recognition method, which comprises the steps that an initial coal flow blocking state recognition model is constructed, and the initial coal flow blocking state recognition model is constructed based on a YOLOv8n target detection algorithm, a StarNet backbone network, a context star fusion module and a weight sharing detection head; the StarNet backbone network is used as a backbone of the YOLOv8n target detection algorithm; a target data set is obtained, the target data set comprises a plurality of sample coal flow images, and the sample coal flow images comprise scraper coal flow blockage information; training the initial coal flow blockage state recognition model based on the target data set to obtain a target coal flow blockage state recognition model; and performing coal flow blockage state identification on the target scraper based on the target coal flow blockage state identification model. Therefore, the precision and the real-time performance of the coal flow blockage detection method can be effectively improved by optimizing a model structure and introducing an innovative module.

Description

Technical Field

[0001] The present disclosure relates to the technical field of image recognition, and in particular to a method for identifying a blockage state of coal flow in a scraper conveyor at a fully mechanized mining face in a coal mine. Background Art

[0002] Coal flow refers to the flow state of coal on the conveyor belt under the action of equipment such as scrapers. Usually, coal flow is composed of a large number of smaller coal particles or coal blocks. These coal particles are relatively loose during transportation and can move on the conveyor belt like a fluid. Coal flow blockage is a common safety hazard in coal mine fully mechanized mining working faces. Especially during the mining process, abnormal blockage of coal flow may cause scraper breakage, coal mining machine failure, poor ventilation, and even more serious mine accidents.

[0003] In the related art, when identifying the blockage state of coal flow in a scraper conveyor, the detection method has low accuracy and poor real-time performance. Summary of the invention

[0004] The present disclosure aims to solve one of the technical problems in the related art at least to some extent.

[0005] To this end, the purpose of the present invention is to propose a method for identifying the coal flow blockage state of a scraper conveyor in a fully mechanized mining face of a coal mine, which can effectively improve the accuracy and real-time performance of the coal flow blockage detection method by optimizing the model structure and introducing innovative modules.

[0006] To achieve the above-mentioned purpose, the method for identifying the blockage state of coal flow in a scraper of a fully mechanized coal mining face in the embodiment of the present disclosure includes: Constructing an initial coal flow blockage state recognition model, wherein the initial coal flow blockage state recognition model is constructed based on the YOLOv8n target detection algorithm, the StarNet backbone network, the context star fusion module and the weight sharing detection head, and the StarNet backbone network serves as the backbone of the YOLOv8n target detection algorithm; Acquire a target data set, wherein the target data set includes a plurality of sample coal flow images, and the sample coal flow images include scraper conveyor coal flow blockage information; Training the initial coal flow blockage state recognition model based on the target data set to obtain a target coal flow blockage state recognition model; The coal flow blockage state of the target scraper is identified based on the target coal flow blockage state identification model.

[0007] The present disclosure provides a method for identifying the coal flow blockage state of a scraper of a fully mechanized coal mining face, which is constructed by constructing an initial coal flow blockage state recognition model, wherein the initial coal flow blockage state recognition model is constructed based on a YOLOv8n target detection algorithm, a StarNet backbone network, a contextual star fusion module, and a weight sharing detection head, and the StarNet backbone network serves as the backbone of the YOLOv8n target detection algorithm; obtaining a target data set, wherein the target data set includes a plurality of sample coal flow images, and the sample coal flow images contain scraper coal flow blockage information; training an initial coal flow blockage state recognition model based on the target data set to obtain a target coal flow blockage state recognition model; and identifying the coal flow blockage state of a target scraper based on the target coal flow blockage state recognition model. Thus, the accuracy and real-time performance of the coal flow blockage detection method can be effectively improved by optimizing the model structure and introducing innovative modules.

[0008] Additional aspects and advantages of the present disclosure will be given in part in the following description and in part will be obvious from the following description or learned through practice of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] The above and / or additional aspects and advantages of the present disclosure will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which: Figure 1 It is a flow chart of a method for identifying a coal flow blockage state of a scraper of a fully mechanized mining face in a coal mine proposed in an embodiment of the present disclosure; Figure 2 It is a flow chart of a method for identifying a coal flow blockage state of a scraper of a fully mechanized mining face in a coal mine proposed in another embodiment of the present disclosure; Figure 3 It is a flow chart of a method for identifying a coal flow blockage state of a scraper of a fully mechanized mining face in a coal mine proposed in another embodiment of the present disclosure; Figure 4 It is a schematic diagram of the principle of the method for identifying the blockage state of coal flow of a scraper of a fully-mechanized mining face in a coal mine proposed in the present disclosure; Figure 5 is a schematic diagram of a network structure of a target detection method proposed in the present disclosure; Figure 6 Schematic diagram of the contextual anchor attention mechanism module proposed in the present disclosure. DETAILED DESCRIPTION

[0010] Embodiments of the present disclosure are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present disclosure, and are not to be construed as limitations of the present disclosure. On the contrary, the embodiments of the present disclosure include all changes, modifications, and equivalents that fall within the spirit and connotation of the appended claims.

[0011] Figure 1 It is a flow chart of a method for identifying the blockage state of coal flow in a scraper conveyor of a fully mechanized mining face in a coal mine proposed in an embodiment of the present disclosure.

[0012] It should be noted that the executor of the method for identifying the coal flow blockage state of a scraper conveyor in a coal mine comprehensive mining working face of this embodiment is a device for identifying the coal flow blockage state of a scraper conveyor in a coal mine comprehensive mining working face, which can be implemented by software and / or hardware.

[0013] like Figure 1 As shown, the method for identifying the blockage state of the coal flow of the scraper of the fully mechanized mining working face of a coal mine comprises: S101: construct an initial coal flow blockage state recognition model, wherein the initial coal flow blockage state recognition model is constructed based on the YOLOv8n target detection algorithm, the StarNet backbone network, the context star fusion module and the weight sharing detection head, and the StarNet backbone network serves as the backbone of the YOLOv8n target detection algorithm.

[0014] The initial coal flow blockage state recognition model refers to a model pre-built in the embodiment of the present disclosure for identifying the coal flow blockage state. The initial coal flow blockage state recognition model can be constructed based on the YOLOv8n target detection algorithm, using the StarNet backbone network as the backbone network of the YOLOv8n target detection algorithm, and adding a contextual star fusion module and a weight sharing detection head.

[0015] Among them, the Contextual Star Fusion Module (CSFM) is an innovative design of StarNet in the target detection task, which aims to enhance the model's global understanding of targets in the image. The main function of this module is to effectively fuse local information with global context information to improve detection performance.

[0016] Among them, Weight Shared Detection Head (WSD Head) is a design method in the target detection model, which aims to improve the parameter efficiency of the detection head part and reduce the computational overhead. Its core idea is to reduce the redundant parameters in the model by sharing the weights of multiple detection heads, while still being able to perform efficient detection on multiple tasks or multiple different target categories.

[0017] In the embodiment of the present disclosure, when the initial coal flow blockage state identification model is constructed, a reliable execution tool can be provided for the subsequent coal flow blockage state identification.

[0018] S102: Acquire a target data set, wherein the target data set includes a plurality of sample coal flow images, and the sample coal flow images contain scraper conveyor coal flow blockage information.

[0019] The target data set may refer to a data set obtained in the embodiment of the present disclosure and used to train the initial coal flow blockage state recognition model.

[0020] In the disclosed embodiment, when acquiring the target data set, the coal conveying video of the scraper may be captured by a working face camera, and a first data set may be obtained after frame extraction processing, and the first data set may be merged with the public data set to obtain the target data set.

[0021] The sample coal flow image may refer to an image containing information related to the blockage of the scraper coal flow.

[0022] That is to say, in the embodiment of the present disclosure, images containing coal flow blockage information can be obtained to construct a target data set, thereby providing reliable data support for the subsequent training of the initial coal flow blockage state recognition model.

[0023] S103: Training an initial coal flow blockage state recognition model based on the target data set to obtain a target coal flow blockage state recognition model.

[0024] The target coal flow blockage state recognition model refers to the coal flow blockage state recognition model obtained by training the initial coal flow blockage state recognition model through the target data set.

[0025] In the embodiment of the present disclosure, when training the initial coal flow blockage state recognition model based on the target data set, the target data set can be divided into a training set, a validation set and a test set according to a certain ratio, and then the processed data set is input into the initial coal flow blockage state recognition model for training and the model weight file with the highest accuracy is saved.

[0026] In the disclosed embodiment, when the initial coal flow blockage state recognition model is trained based on the target data set to obtain the target coal flow blockage state recognition model, a reliable execution tool can be provided for the subsequent identification of the coal flow blockage state of the target scraper.

[0027] S104: Identify the coal flow blockage state of the target scraper based on the target coal flow blockage state identification model.

[0028] Among them, the target scraper refers to the scraper for which the coal flow blockage state is to be identified in the embodiment of the present disclosure.

[0029] That is to say, in the embodiment of the present disclosure, after training the initial coal flow blockage state recognition model based on the target data set to obtain the target coal flow blockage state recognition model, the coal flow blockage state of the target scraper can be identified based on the target coal flow blockage state recognition model, thereby effectively improving the efficiency and intelligence of coal flow blockage state recognition.

[0030] In this embodiment, an initial coal flow blockage state recognition model is constructed, wherein the initial coal flow blockage state recognition model is constructed based on the YOLOv8n target detection algorithm, the StarNet backbone network, the context star fusion module and the weight sharing detection head, and the StarNet backbone network serves as the backbone of the YOLOv8n target detection algorithm; a target data set is obtained, wherein the target data set includes a plurality of sample coal flow images, and the sample coal flow images contain scraper coal flow blockage information; an initial coal flow blockage state recognition model is trained based on the target data set to obtain a target coal flow blockage state recognition model; and the target scraper is subjected to coal flow blockage state recognition based on the target coal flow blockage state recognition model. Thus, the accuracy and real-time performance of the coal flow blockage detection method can be effectively improved by optimizing the model structure and introducing innovative modules.

[0031] Figure 2 It is a flow chart of a method for identifying the blockage state of coal flow in a scraper conveyor of a coal mine comprehensive mining face proposed in another embodiment of the present disclosure.

[0032] like Figure 2 As shown, the method for identifying the blockage state of the coal flow of the scraper of the fully mechanized mining working face of a coal mine comprises: S201: construct an initial coal flow blockage state recognition model, wherein the initial coal flow blockage state recognition model is constructed based on the YOLOv8n target detection algorithm, the StarNet backbone network, the context star fusion module and the weight sharing detection head, and the StarNet backbone network serves as the backbone of the YOLOv8n target detection algorithm.

[0033] S202: Acquire a target data set, wherein the target data set includes a plurality of sample coal flow images, and the sample coal flow images contain scraper conveyor coal flow blockage information.

[0034] S203: Training an initial coal flow blockage state recognition model based on the target data set to obtain a target coal flow blockage state recognition model.

[0035] The description of S201 - S203 can be specifically referred to the above embodiment, which will not be repeated here.

[0036] S204: Acquire a real-time coal flow image of the target scraper based on a preset time interval.

[0037] The preset time interval refers to the time interval of the target scraper corresponding to the coal flow image preset in the present disclosure, which can be, for example, 1 second, 10 seconds, etc., without limitation.

[0038] The real-time coal flow image refers to the coal flow image collected in the actual production operation environment.

[0039] That is to say, in the embodiment of the present disclosure, the real-time coal flow image of the target scraper can be acquired based on the preset time interval, thereby providing reliable data support for the subsequent identification of the coal flow blockage state.

[0040] S205: Perform recognition processing on the real-time coal flow image based on the target coal flow blockage state recognition model to obtain a coal flow blockage state recognition result.

[0041] The coal flow blockage state identification result can be used to indicate whether the target scraper has coal flow blockage, or can further indicate the blockage degree of the coal flow.

[0042] That is, in the embodiment of the present disclosure, after the initial coal flow blockage state recognition model is trained based on the target data set to obtain the target coal flow blockage state recognition model, the real-time coal flow image of the target scraper can be obtained based on the preset time interval; the real-time coal flow image is recognized and processed based on the target coal flow blockage state recognition model to obtain the coal flow blockage state recognition result. In this way, the real-time monitoring of the coal flow state can be realized, thereby effectively improving the timeliness and reliability of the obtained coal flow blockage state recognition.

[0043] S206: When the coal flow blockage state identification result indicates that the coal flow in the real-time coal flow image is in a blockage state, an alarm message is generated according to the real-time coal flow image, and a coal flow blockage warning is issued based on the alarm message.

[0044] The alarm information can be used for early warning of coal flow blockage. For example, the alarm information can be used to indicate whether the target scraper has coal flow blockage, or can also be used to indicate the degree of coal flow blockage, etc., without limitation.

[0045] Optionally, in some embodiments, when generating alarm information based on the real-time coal flow image, the volume of the coal pile where the target scraper is blocked can be determined based on the real-time coal flow image, and the alarm information can be generated based on the volume of the coal pile. Thus, the alarm information obtained can accurately indicate the blockage degree of the target scraper, so as to effectively improve the practicality of the alarm information.

[0046] Optionally, in some embodiments, when generating alarm information based on the real-time coal flow image, the position of the coal pile where the target scraper is blocked can be determined based on the real-time coal flow image, and the alarm information can be generated based on the coal pile position. Thus, the alarm information obtained can accurately indicate the position of the coal pile in the blocked state, so as to realize the rapid positioning of the blocked coal flow, thereby effectively improving the efficiency of coal flow blockage treatment.

[0047] Optionally, in some embodiments, when performing coal flow blockage warning based on alarm information, at least one of the following may be included: Light warning based on alarm information; Voice warning based on alarm information; A coal flow blockage status chart is generated based on the alarm information, and a visual early warning is performed based on the coal flow blockage status chart.

[0048] Therefore, a personalized approach can be adopted to implement early warning of coal flow blockage, so as to effectively improve the early warning effect.

[0049] The coal flow blockage state chart refers to a chart describing the coal flow blockage state, for example, a line chart.

[0050] That is, in the embodiment of the present disclosure, when the coal flow blockage state recognition result indicates that the coal flow is in a blocked state in the real-time coal flow image, alarm information can be generated according to the real-time coal flow image, and a coal flow blockage warning can be performed based on the alarm information. Thus, corresponding alarm information can be accurately generated based on the real-time coal flow image, and a coal flow blockage warning can be performed based on the alarm information.

[0051] In this embodiment, the real-time coal flow image of the target scraper is obtained based on a preset time interval; the real-time coal flow image is identified and processed based on the target coal flow blockage state recognition model to obtain the coal flow blockage state recognition result. In this way, real-time monitoring of the coal flow state can be achieved, thereby effectively improving the timeliness and reliability of the obtained coal flow blockage state recognition. When the coal flow blockage state recognition result indicates that the real-time coal flow image has a coal flow in a blocked state, alarm information is generated according to the real-time coal flow image, and a coal flow blockage warning is performed based on the alarm information. In this way, the corresponding alarm information can be accurately generated based on the real-time coal flow image, and a coal flow blockage warning is performed based on the alarm information.

[0052] Figure 3It is a flow chart of a method for identifying the blockage state of coal flow in a scraper conveyor of a coal mine comprehensive mining face proposed in another embodiment of the present disclosure.

[0053] like Figure 3 As shown, the method for identifying the blockage state of the coal flow of the scraper of the fully mechanized mining working face of a coal mine comprises: S301: construct an initial coal flow blockage state recognition model, wherein the initial coal flow blockage state recognition model is constructed based on the YOLOv8n target detection algorithm, the StarNet backbone network, the context star fusion module and the weight sharing detection head, and the StarNet backbone network serves as the backbone of the YOLOv8n target detection algorithm.

[0054] S302: Acquire a target data set, wherein the target data set includes a plurality of sample coal flow images, and the sample coal flow images contain scraper conveyor coal flow blockage information.

[0055] S303: Training an initial coal flow blockage state recognition model based on the target data set to obtain a target coal flow blockage state recognition model.

[0056] S304: Acquire a real-time coal flow image of the target scraper based on a preset time interval.

[0057] S305: Perform recognition processing on the real-time coal flow image based on the target coal flow blockage state recognition model to obtain a coal flow blockage state recognition result.

[0058] The description of S301 - S305 can be specifically referred to the above embodiment, which will not be repeated here.

[0059] S306: When the coal flow blockage state identification result indicates that the coal flow in the real-time coal flow image is in a blockage state, a preset control instruction is sent to the target scraper, wherein the preset control instruction is used to instruct the target scraper to stop running.

[0060] The preset control instruction may refer to a pre-configured control instruction for instructing the target scraper to stop running.

[0061] In this embodiment, when the coal flow blocking state recognition result indicates that the coal flow in the real-time coal flow image is in a blocking state, a preset control instruction is sent to the target scraper, wherein the preset control instruction is used to instruct the target scraper to stop running. Thus, when the coal flow is in a blocking state, the target scraper can be promptly instructed to stop running based on the preset control instruction, so as to effectively improve the safety and robustness of the operation process of the target scraper.

[0062] Based on the above embodiments, Figure 4 As shown, Figure 4The present invention is a schematic diagram of the principle of the method for identifying the blockage state of the coal flow of the scraper of the fully-mechanized working face of a coal mine proposed in the present invention. First, the coal conveying video of the scraper is collected by the working face camera, and the frame is extracted and merged with the data set. The merged data set is manually visually annotated, and only the area where the coal flow is blocked is annotated, and the training set, the verification set and the test set are divided according to a certain ratio; then, the processed data set is input into Star-YOLOv8 for training, and the model weight file with the highest accuracy is saved. Finally, the weight file, the inference algorithm and the post-processing algorithm are deployed to the back-end service of the mine safety patrol software, and the service is deployed to the Linux system server, and the algorithm and the real-time video stream provided by the working face camera are bound by sending a configuration file to the back-end service. It can also be deployed on the edge computing server. Each frame of data is processed in real time. If the coal flow is blocked, a response instruction is immediately sent to the PLC, and the PLC sends an instruction to control the relay to light up, and the relay sends an instruction information to control the buzzer alarm, and the operation of the scraper can be shut down. The coal flow blockage location, pictures and video clips are uploaded to the server for storage, and the video inference stream is displayed in real time on the front end of the mine safety patrol software. The alarm information is displayed in real time on the early warning interface through visualization forms such as charts.

[0063] This paper proposes a lightweight target detection method (Star-YOLOv8) based on YOLOv8n to improve the computational efficiency and detection accuracy of the model. Figure 5 As shown, Figure 5 Schematic diagram of the network structure of the target detection method proposed in the present disclosure.

[0064] First, the present invention replaces the backbone network of YOLOv8 with StarNet. StarNet is an efficient neural network architecture that introduces "star operations" (i.e., element-wise multiplication) to perform feature mapping, thus avoiding the computational complexity brought by traditional convolution operations. While maintaining a low computational burden, StarNet can effectively enhance feature extraction capabilities, especially for multi-scale feature fusion, thereby improving the performance of the model in complex coal flow blockage detection scenarios.

[0065] Secondly, in order to further improve the feature expression capability, this paper designs a contextual star fusion module (CSFM), such as Figure 5As shown in the figure. This module combines the star operation in StarBlock and further enhances the model's context perception and focus on the central area through multi-dimensional feature fusion and context anchor attention mechanism (CAA). This design not only enables the model to better capture the characteristics of coal flow blockage state on a lightweight basis, but also effectively reduces the computational overhead of multi-level feature fusion in traditional models.

[0066] Finally, the present disclosure proposes a weighted shared detection head (WSDHead), such as Figure 5 Detection head. This module reduces the amount of calculation by uniformly processing feature maps and performing convolution operations, while improving the model's real-time detection capabilities. The weight sharing mechanism can effectively avoid repeated calculations, making the model more efficient and adaptable to the needs of real-time applications.

[0067] like Figure 5 As shown in the figure, the overall framework of the model includes a feature extraction backbone network (BackBone), a neck module (Neck) for enhanced feature expression, and a lightweight shared weight detection head (WSD).

[0068] The overall framework of the model is based on StarNet as the backbone network (BackBone). StarNet is a new type of neural network architecture that implements high-dimensional nonlinear feature mapping through star operations (i.e. element-level multiplication). It outperforms many designs without complex designs or carefully selected hyperparameters, and is lighter than the backbone network of YOLOv8n.

[0069] The backbone (BackBone) network of StarNet includes P1, P2, P3, P4 and P5, a total of 5 feature extraction modules, P1 has only 1 convolutional layer; P2, P3, P4 and P5 each have n star blocks, and StarNets of different sizes are constructed by changing the number of blocks and the number of input embedding channels. After the P5 module, the network is connected to a SPPF (Spatial PyramidPooling - Fast) layer, which improves the inference speed by optimizing the feature map pooling process and reducing the amount of calculation. Then the feature maps output by the P3, P4 and SPPF layers are fused and spliced ​​at the neck ( Figure 5 "C" is used in the code to represent the concatenation operation, i.e. Concat).

[0070] In order to enhance the model's ability to express features of multi-scale targets and complex scenes, the YOLOv8 bottleneck structure is improved. This paper designs a contextual star fusion module (CSFM) and replaces the C2f (Faster Implementation of CSP Bottleneck with 2 convolutions) module of YOLOv8 with CSFM. Figure 5 As shown in the figure, F is used to represent the feature map. The feature map F1 output by the SPPF layer is upsampled to obtain F2. F2 is concatenated with the output feature map of the P4 layer, and the result F3 is input into the context star fusion module (CSFM) for processing. The output feature map F4 is upsampled once and then concatenated with the feature map output by the P3 layer to obtain F6. F6 is input into the CSFM module for processing to obtain F7. F7 is convolved once to obtain the feature map F8. F8 and F4 are concatenated to obtain F9. F9 is processed by the CSFM module and convolved once to obtain F 10 , F 10 It is concatenated with the feature map output by the SPPF layer to obtain F 11 , F 11 After another CSFM, we get F 12 , F7, F 13 and F 12 Input the weighted shared detection head (WSD) to get the final detection result.

[0071] The weighted shared detection head (WSD) simplifies the calculation process by optimizing the structure, further reducing the number of model parameters and improving real-time detection capabilities.

[0072] For the complex scenario of coal flow blockage identification, the model is required to enhance the ability to focus on important features and the extraction accuracy while ensuring lightweight and real-time performance. To this end, this paper designs a contextual star fusion module (CSFM). By introducing the contextual anchor attention mechanism (CAA) in PKINet into the StarBlock structure of StarNet, the new structure is referred to as the StarBlock-CAA module, and the module replaces the Bottleneck part of the C3 module in YOLOv5. On the basis of maintaining lightweight, the model enhances the focus on multi-scale features and improves the accuracy and real-time performance in coal flow blockage identification.

[0073] CAA is an attention mechanism designed to enhance the feature representation capability of neural networks and is suitable for tasks that require rich contextual information. It not only focuses on the global context, but also strengthens the features of the central area of ​​the image through the attention mechanism, so that the model pays more attention to features related to coal flow blockage and improves the recognition accuracy of the target area. Compared with traditional large-kernel convolution, the CAA module has less computation and parameters, and the model is more lightweight.

[0074] The CAA module weights the input feature map through the channel attention mechanism. Figure 6 As shown, Figure 6 is a schematic diagram of the context anchor attention mechanism module proposed in the present disclosure. First, the input feature map conduct Average pooling (AvgPool), after one-dimensional convolution operation, obtains Then input horizontal convolution ,get , after vertical convolution get Finally, after one-dimensional convolution and Sigmoid activation function, the channel attention factor is obtained The final output of the CAA module is:

[0075] here, It is the channel attention weight obtained through the above steps, which is multiplied with the original feature map to complete the channel weighted operation.

[0076] Adding the CAA module after the element-wise multiplication step of StarBlock can further improve the model's detection performance for objects in images without significantly increasing the computational burden.

[0077] like Figure 5 As shown in the figure, the StarBlock-CAA module first processes the input feature map through a depthwise convolution (DW-Conv). Subsequently, the processed feature maps are input into two fully connected layers (FC) respectively, and the outputs of the two branches are multiplied element by element. Next, the multiplication result is enhanced by the contextual anchor attention mechanism (CAA) to capture the contextual information in the feature map. After that, the fully connected layer (FC) is applied to the output of CAA again for feature transformation, and the spatial features are further extracted through another depthwise separable convolution (DW-Conv). Finally, the results of all the above operations are concatenated (Concat) to form the final output of the StarBlock-CAA module.

[0078] The C3 module in YOLOv5 is a module that contains multiple convolutional layers, feature fusion, and information enhancement. It is composed of multiple ConvBNSiLU combination units and a bottleneck layer Bottleneck. Among them, ConvBNSiLU consists of a convolution layer (Convolution), a batch normalization layer (Batch Normalization), and a SiLU (Sigmoid Linear Unit) activation function layer, which aims to improve the efficiency and stability of feature extraction. This study uses StarBlock-CAA to replace Bottleneck in the C3 module to form a CSFM lightweight architecture.

[0079] This design can significantly improve the performance of the model. CSFM uses the core element multiplication operation of StarNet to effectively map the input features into a high-dimensional nonlinear feature space, enhancing the model's ability to express features. At the same time, the introduced CAA module design is relatively lightweight and can capture rich contextual information without significantly increasing the computational burden, further improving feature expression.

[0080] This paper designs a lightweight weight-sharing detection head (WSD). By sharing convolution modules to uniformly process feature maps, it not only reduces the number of parameters, but also effectively integrates spatial context information of different scales. This sharing mechanism avoids the redundancy caused by independent design of each layer, improves computational efficiency and overall consistency of the model.

[0081] according to Figure 5 The detection head network structure shown in the figure, P3, P4 and P5 represent the feature maps after neck convolution and feature fusion, representing the feature maps from smaller scale (P3) to larger scale (P5). Each feature map is first processed by a 1x1 ConvGN unit (including 2D convolution Conv2d, group normalization GroupNorm and SiLU activation function). Then, the output results of these three branches are input into two 3x3 ConvGN units for further processing. In this way, the network can use feature maps of different scales for target detection, thereby outputting the detection box coordinates of the target and the category of the target in the box.

[0082] The following is the specific process of WSD.

[0083] First, extract channel features: (1) in, Represents the input feature map corresponding to each scale, which are P3, P4, P5, Represents a ConvGN combination unit (such as Figure 5Each scale feature map P3, P4, P5 will produce As the feature after convolution. Here and in the following formulas, .

[0084] Next, the context information of feature maps of different scales is further extracted through the shared convolution module: (2) in, Two consecutive Convolution operation is performed on each feature map.

[0085] Each scale feature map predicts the bounding box and category separately. Bounding box regression branch (each scale feature map outputs a bounding box prediction): (3) in, is the discretized maximum value of bounding box regression (the granularity of the distribution), Represents the coordinates of the predicted target box.

[0086] Category prediction branch: (4) in, is the number of categories, Represents the category prediction of each scale feature map.

[0087] The predictions at each scale are rescaled to obtain the final bounding box regression output: (5) Bounding box predictions at each scale All are decoded by the DFL (distributed focus loss) module. The DFL module performs distribution regression on the bounding box regression value: (6) The detailed steps of the DFL module are as follows, 1. Reshape the distribution value: (7) Where B is the batch size, 4 means x, y, w, h, is the discretization granularity of the distribution, and A is the number of anchor points.

[0088] 2. Soft maximization, applying the softmax operation to the distribution: (8) 3. Weighted summation, weighted summation of the discretized regression distribution: (9) The resulting decoded bounding box is: (10) Decode the bounding box coordinates, and decode the bounding box coordinates in combination with the anchor points. dist2bbox represents the distance decoding function, anchors represents the anchor points, and strides represents the scale factor of each detection layer: (11) Perform category prediction. The category prediction is activated by Sigmoid to obtain the final category confidence: (12) By concatenating the bounding boxes and category predictions of all scales, we get the final output: (13) The lightweight real-time detection method of coal flow blockage state based on Star-YOLOv8 of the present invention has significant beneficial effects. First, by introducing the innovative contextual star fusion module (CSFM), the accuracy of coal flow blockage detection is significantly improved, and the coal flow blockage state can be accurately identified and distinguished in a complex coal mine environment, effectively reducing the situation of false detection and missed detection. Secondly, the adoption of the lightweight YOLOv8 model enables the detection process to run efficiently on edge devices while ensuring high accuracy, greatly improving the real-time performance of the detection and meeting the demand for rapid response at the coal mine production site. At the same time, due to the reduced dependence on high-performance computing hardware, computing resources and energy consumption are effectively saved, providing a more economical and environmentally friendly solution for coal mine production. The automated detection function of the system simplifies the operation process, reduces manual intervention, reduces the difficulty of operation, and improves work efficiency. In addition, the present invention can detect coal flow blockage problems in a timely and accurate manner, thereby providing real-time warnings for miners, avoiding equipment damage and production stoppage accidents, effectively ensuring coal mine production safety, reducing production interruptions and equipment maintenance costs caused by blockage, and improving overall production efficiency.

[0089] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. The present disclosure is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The description and examples are to be considered exemplary only, and the true scope and spirit of the present disclosure are indicated by the following claims.

[0090] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

[0091] It should be noted that, in the description of the present disclosure, the terms "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance. In addition, in the description of the present disclosure, unless otherwise specified, the meaning of "plurality" is two or more.

[0092] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code that includes one or more executable instructions for implementing the steps of a specific logical function or process, and the scope of the preferred embodiments of the present disclosure includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present disclosure belong.

[0093] It should be understood that the various parts of the present disclosure may be implemented in hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods may be implemented in software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof may be used to implement: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0094] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0095] In addition, each functional unit in each embodiment of the present disclosure may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0096] The storage medium mentioned above can be a read-only memory, a magnetic disk or an optical disk, etc.

[0097] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present disclosure. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0098] Although the embodiments of the present disclosure have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present disclosure. A person skilled in the art may change, modify, replace and vary the above embodiments within the scope of the present disclosure.

Claims

1. A method for identifying the blockage state of coal flow in a scraper of a fully mechanized mining face in a coal mine, characterized in that: include: Constructing an initial coal flow blockage state recognition model, wherein the initial coal flow blockage state recognition model is constructed based on the YOLOv8n target detection algorithm, the StarNet backbone network, the context star fusion module and the weight sharing detection head, and the StarNet backbone network serves as the backbone of the YOLOv8n target detection algorithm; Acquire a target data set, wherein the target data set includes a plurality of sample coal flow images, and the sample coal flow images include scraper conveyor coal flow blockage information; Training the initial coal flow blockage state recognition model based on the target data set to obtain a target coal flow blockage state recognition model; The coal flow blockage state of the target scraper is identified based on the target coal flow blockage state identification model.

2. The method according to claim 1, characterized in that The identifying the coal flow blockage state of the target scraper based on the target coal flow blockage state identification model includes: Acquire a real-time coal flow image of the target scraper based on a preset time interval; The real-time coal flow image is identified and processed based on the target coal flow blockage state identification model to obtain a coal flow blockage state identification result.

3. The method according to claim 2, characterized in that The method further comprises: When the coal flow blockage state identification result indicates that the coal flow in the real-time coal flow image is in a blockage state, alarm information is generated according to the real-time coal flow image, and a coal flow blockage warning is issued based on the alarm information.

4. The method according to claim 3, characterized in that The generating of alarm information according to the real-time coal flow image comprises: The volume of the coal pile where the target scraper is in a blocked state is determined according to the real-time coal flow image, and the alarm information is generated according to the volume of the coal pile.

5. The method according to claim 3, characterized in that The generating of alarm information according to the real-time coal flow image comprises: The coal pile position where the target scraper is in a blocked state is determined according to the real-time coal flow image, and the alarm information is generated according to the coal pile position.

6. The method according to claim 3, characterized in that The coal flow blockage warning based on the alarm information includes at least one of the following: Performing a lighting warning based on the alarm information; Producing a voice warning based on the alarm information; A coal flow blockage status chart is generated based on the alarm information, and a visual early warning is performed based on the coal flow blockage status chart.

7. The method according to claim 2, characterized in that The method further comprises: When the coal flow blockage state identification result indicates that the coal flow in the real-time coal flow image is in a blockage state, a preset control instruction is sent to the target scraper, wherein the preset control instruction is used to instruct the target scraper to stop running.

Images