Large coal intelligent identification system

Abstract

The invention belongs to the technical field of large coal recognition, and particularly relates to an intelligent large coal recognition system which comprises a data acquisition module used for acquiring coal images, three-dimensional space information of coal and physical parameters of the coal; and the fault diagnosis module is used for monitoring the operation state of each acquisition device in real time so as to quickly position a fault point by analyzing the working parameters and data transmission conditions of the devices. According to the invention, coal images and data can be continuously processed in real time in the whole process of coal production, processing and transportation; compared with manual work, the system has the advantages that fatigue, distraction and the like are avoided, a plurality of images can be processed per second, a processing scene of dozens of tons or even hundreds of tons of coal per hour can be quickly and accurately identified, coal processing flow overstocking is effectively avoided, and the overall production efficiency is remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of bulk coal identification, and in particular to an intelligent bulk coal identification system. Background Art

[0002] In coal production, processing, and transportation, the size of coal lumps is a critical parameter, directly affecting coal quality assessment, processing efficiency, and subsequent sales and use. Traditionally, the identification of large lumps of coal has relied primarily on manual labor, a method with numerous drawbacks. Specifically, the following drawbacks exist:

[0003] 1. Inefficient manual identification: Coal production sites, such as coal mines and coal preparation plants, are plagued by numerous coal lumps. Manual identification of large lumps requires workers to observe and identify them for extended periods of time, resulting in high workload and slow identification speeds. For example, a large coal preparation plant may process tens or even hundreds of tons of coal per hour. Relying on manual identification of large lumps individually would be difficult to meet production efficiency requirements and could easily lead to backlogs in the coal processing process, impacting overall production progress.

[0004] 2. Low manual recognition accuracy: Manual identification of large lumps of coal is susceptible to various factors, resulting in unstable results. A worker's experience, fatigue, and concentration can all affect recognition results. For example, a new employee may lack experience and misjudge the size of large lumps of coal; while those who have worked for a long time may make mistakes due to fatigue. Inaccurate identification of large lumps of coal during coal quality testing can lead to deviations in coal quality assessments, affecting coal sales prices and market competitiveness.

[0005] 3. Manual identification poses safety risks: Various safety risks exist in coal production environments, such as underground in coal mines and near coal transport belts. Manual identification of large lumps of coal in these environments requires workers to be close to operating equipment or in hazardous areas, making accidents more likely. For example, underground in coal mines, workers must identify large lumps of coal in dusty, dimly lit environments. Not only are the working conditions difficult, but there are also safety hazards such as gas explosions and roof collapse, threatening workers' lives.

[0006] Based on the above, a large coal intelligent identification system is invented. Summary of the Invention

[0007] To solve the above technical problems, according to one aspect of the present invention, the present invention provides the following technical solutions:

[0008] A large coal intelligent identification system, comprising:

[0009] A data acquisition module is used to collect coal images, three-dimensional spatial information of coal, and physical parameters of coal;

[0010] The fault diagnosis module is used to monitor the operating status of each acquisition device in real time, so as to quickly locate the fault point by analyzing the device's working parameters and data transmission status;

[0011] The data transmission module is used to transmit the data collected by the data acquisition module. To ensure the stability and reliability of data transmission, data encryption and data verification technologies will be used;

[0012] The data processing module is used to process and analyze the collected data, including image preprocessing, feature extraction, model training, and recognition and classification;

[0013] The application module is used to provide users with various application functions of the system.

[0014] As a preferred solution of the intelligent identification system for large coal lumps described in the present invention, the fault diagnosis module includes:

[0015] The equipment status information collection module is used to establish communication connections with various devices in the data acquisition module to obtain key parameter information of equipment operation in real time;

[0016] The data preprocessing module is used to preprocess the received raw data. It first uses a filtering algorithm to remove noise interference in the data, and then uses interpolation or statistical models to complete the data based on the historical operation data of the device and the status information of adjacent devices to ensure the accuracy and completeness of the subsequent analysis data.

[0017] As a preferred solution of the intelligent identification system for large coal lumps described in the present invention, the fault diagnosis module further includes:

[0018] The fault threshold setting and judgment module is used to first set a reasonable threshold range for each monitoring parameter based on the technical parameters of the equipment and historical data of normal operation. It then compares the real-time collected and pre-processed data with the set threshold. If a parameter is found to exceed the threshold range, the fault detection process is immediately triggered.

[0019] The fault type location module is used to accurately locate the fault type after detecting abnormal parameters through fault tree analysis, neural network intelligent algorithms, combined with the changes in various parameters of the equipment and the fault knowledge base.

[0020] As a preferred solution of the intelligent identification system for large coal lumps described in the present invention, the fault diagnosis module further includes:

[0021] The fault alarm and information push module is used to immediately send out an alarm signal after determining the fault type and location, and to notify the fault on site through the sound and light alarm device. At the same time, detailed fault information is sent to equipment maintenance personnel and relevant management personnel via SMS, email or system message push to ensure that the responsible person is notified as soon as possible.

[0022] The fault repair and feedback module is used to enable maintenance personnel to perform on-site repairs based on the information provided after receiving alarm information. After the repair is completed, it can provide feedback on the fault repair status and re-collect equipment operating parameters to confirm that the equipment has returned to normal operating status. At the same time, it updates the fault knowledge base to provide richer experience data for subsequent fault diagnosis.

[0023] As a preferred solution of the intelligent identification system for large coal lumps described in the present invention, the data processing module includes:

[0024] The data receiving and diversion module is used to first receive the data collected by the data acquisition module, and then divert it according to the data type and subsequent processing requirements;

[0025] Image preprocessing module, used to perform denoising, enhancement and segmentation on the collected images;

[0026] The feature extraction module is used to first extract the size, shape and texture of the segmented coal block image, and then perform statistical feature extraction on the physical parameter data;

[0027] The model training module is used to first label and segment the coal image data, then select the appropriate deep learning model and perform optimization and evaluation;

[0028] The recognition and classification module is used to input the real-time coal image data and physical parameter data that have undergone preprocessing and feature extraction into the trained large-lump coal recognition model, so that the model can analyze and judge the input data based on the learned feature patterns and classification rules, identify the large lumps of coal, and mark the location and size information of the large lumps of coal; at the same time, the recognition results are transmitted to the application module for users to conduct real-time monitoring, statistical analysis and production decision-making.

[0029] As a preferred solution of the intelligent identification system for large coal lumps described in the present invention, the specific steps of the image preprocessing module are as follows:

[0030] Step 1: Denoising: Use median filtering and Gaussian filtering algorithms to remove salt and pepper noise and Gaussian noise in the image to improve image clarity;

[0031] Step 2: Image enhancement: Using histogram equalization and adaptive histogram equalization techniques, the image contrast is enhanced to make the edges and details of the coal blocks more prominent, facilitating subsequent feature extraction.

[0032] Step 3: Image segmentation: Use threshold segmentation, region growing, and semantic segmentation algorithms to separate the coal blocks in the coal image from the background and extract the image of a single coal block, laying the foundation for subsequent feature extraction and recognition.

[0033] As a preferred solution of the intelligent identification system for large coal lumps described in the present invention, the specific steps of the feature extraction module are as follows:

[0034] Step 1: Image feature extraction: For the segmented coal block images, deep learning algorithms or traditional image processing methods are used to extract the size, shape, and texture features of the coal blocks. Simultaneously, the 3D spatial information acquired by the LiDAR is combined to calculate the 3D features of the coal blocks, such as their volume.

[0035] Step 2: Extraction of physical parameter features: The physical parameter data collected by the sensor is normalized and standardized to eliminate the impact of dimensional differences between different parameters. At the same time, statistical features are extracted through statistical analysis to provide multi-dimensional information support for the identification of large coal blocks.

[0036] As a preferred solution of the intelligent identification system for large coal lumps described in the present invention, the specific steps of the model training module are as follows:

[0037] Step 1: Data labeling and segmentation: First, collect coal image data, then manually label large pieces of coal to clarify their location and size information; then divide the labeled data into training, validation, and test sets;

[0038] Step 2: Select an appropriate deep learning model: First, select a classic model in the convolutional neural network. Then, optimize and adjust the model structure according to actual needs. Input the training set data into the model. Use the backpropagation algorithm and gradient descent method to continuously adjust the model weight parameters to minimize the error between the model's prediction results and the labeled data.

[0039] Step 3: Model optimization and evaluation: During the training process, the model is evaluated using the validation set data. If overfitting or underfitting occurs, it will be optimized by adjusting the model structure, adding data augmentation operations, and adjusting the learning rate method. Once the model performance on the validation set is stable, the test set is used to conduct a final evaluation of the model to determine the model's generalization ability and reliability.

[0040] Compared with existing technologies:

[0041] 1. Addressing the inefficiency of manual identification: The present invention can process coal images and data in real time and continuously throughout the entire process of coal production, processing, and transportation. Compared with manual labor, the system will not experience fatigue or distraction, and can process multiple images per second. Even in scenarios where tens or even hundreds of tons of coal are processed per hour, it can achieve rapid and accurate identification, effectively avoiding backlogs in the coal processing process and significantly improving overall production efficiency.

[0042] 2. To address the low accuracy of manual identification: This invention uses a deep learning algorithm and trains on a large amount of labeled coal image data to establish an accurate large-lump coal identification model. The algorithm is not affected by subjective factors such as staff experience, emotions, and fatigue, and can stably extract the size, shape, texture and other characteristics of coal blocks, and accurately identify large-lump coal according to preset standards. In the coal quality inspection link, it ensures the reliability of quality assessment and enhances the competitiveness of the company's coal products in the market.

[0043] 3. Addressing the safety risks associated with manual identification: This invention utilizes non-contact image acquisition and analysis, eliminating the need for workers to be physically present in the dusty, dimly lit, and hazardous areas of underground coal mines, or near running equipment such as conveyor belts. The system remotely collects data through high-definition cameras, lidar, and other equipment installed in key locations, effectively preventing accidents such as gas explosions, roof collapses, and mechanical injuries, effectively protecting workers and improving the working environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 It is a schematic diagram of the overall framework of the present invention. DETAILED DESCRIPTION

[0045] To make the objectives, technical solutions and advantages of the present invention more clear, the embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

[0046] The present invention provides a large coal intelligent identification system, please refer to Figure 1 ,include:

[0047] The data acquisition module is used to collect coal images, three-dimensional spatial information, and physical parameters. It includes high-definition cameras, lidars, and sensors. The high-definition cameras are used to capture coal images, the lidars are used to obtain three-dimensional spatial information, and the sensors are used to collect physical parameters such as weight and temperature. These devices will be installed at key locations in coal production, processing, and transportation, such as coal mining faces, coal preparation plant conveyor belts, and coal transport vehicles.

[0048] The fault diagnosis module is used to monitor the operating status of each acquisition device in real time, so as to quickly locate the fault point by analyzing the device's working parameters and data transmission status;

[0049] By setting up a fault diagnosis module, it is convenient for staff to quickly repair the collection equipment when it fails, reducing the impact of equipment failure on the identification of large coal blocks and ensuring the continuous and stable operation of the system.

[0050] The data transmission module is used to transmit the data collected by the data acquisition module. Different transmission methods can be selected according to different application scenarios and data volume, such as wired network transmission, wireless network transmission (such as 4G, 5G), etc. In order to ensure the stability and reliability of data transmission, data encryption and data verification technologies will be adopted;

[0051] The data processing module is used to process and analyze the collected data, including image preprocessing, feature extraction, model training, and recognition and classification;

[0052] The application module is used to provide users with various application functions of the system, including real-time monitoring of large coal lumps, statistical analysis, alarm prompts, production decision support, etc. Users can access the system through the web interface or mobile application to view the identification results of large coal lumps and related data statistics, set alarm thresholds, receive alarm information, etc.

[0053] The fault diagnosis module includes:

[0054] The device status information collection module is used to establish communication connections with various devices in the data acquisition module (such as high-definition cameras, lidars, and sensors) to obtain key operating parameter information of the equipment in real time. For high-definition cameras, it collects information such as lens temperature, image resolution, frame rate, and transmission bandwidth utilization; lidars collect data such as scanning frequency, signal strength, and ranging error; and sensors collect parameters such as the stability of their output data, power supply voltage, and operating temperature. This information is continuously transmitted to the fault diagnosis module through the data transmission channel.

[0055] The data preprocessing module is used to preprocess the received raw data. It first uses a filtering algorithm to remove noise interference in the data. Then, it uses interpolation or statistical models to supplement the data based on the historical operation data of the device and the status information of adjacent devices to ensure the accuracy and completeness of the data for subsequent analysis.

[0056] The fault threshold setting and judgment module is used to first set a reasonable threshold range for each monitoring parameter based on the equipment's technical parameters and historical data of normal operation. For example, an HD camera's lens temperature exceeding 60°C, an image frame rate below the set standard, or a LiDAR ranging error exceeding ±5cm are all considered abnormal conditions. The module then compares the real-time collected and pre-processed data with the set threshold. If a parameter is found to exceed the threshold range, the fault detection process is immediately triggered.

[0057] The fault type location module is used to accurately locate the fault type after detecting abnormal parameters through fault tree analysis, neural network intelligent algorithms, combined with the changes in various equipment parameters and the fault knowledge base;

[0058] The fault alarm and information push module is used to immediately send out an alarm signal after determining the fault type and location, and to notify the fault on site through the sound and light alarm device. At the same time, detailed fault information (including the name of the faulty device, fault type, fault time, possible cause, etc.) is sent to equipment maintenance personnel and relevant management personnel via SMS, email or system message push to ensure that the responsible person is notified as soon as possible;

[0059] The fault repair and feedback module is used to enable maintenance personnel to perform on-site repairs based on the information provided after receiving alarm information. After the repair is completed, it can provide feedback on the fault repair status and re-collect equipment operating parameters to confirm that the equipment has returned to normal operating status. At the same time, it updates the fault knowledge base to provide richer experience data for subsequent fault diagnosis.

[0060] The data processing module includes:

[0061] The data receiving and diversion module is used to first receive the data collected by the data acquisition module, and then divert it according to the data type and subsequent processing requirements, so as to guide the coal image data to the image preprocessing module, and transmit the three-dimensional spatial information and physical parameter data to the corresponding feature extraction module respectively, to ensure that the data enters the subsequent processing flow in an orderly manner;

[0062] Image preprocessing module, used to perform denoising, enhancement and segmentation on the collected images;

[0063] The feature extraction module is used to first extract the size, shape and texture of the segmented coal block image, and then perform statistical feature extraction on the physical parameter data;

[0064] The model training module is used to first label and segment the coal image data, then select the appropriate deep learning model and perform optimization and evaluation;

[0065] The recognition and classification module is used to input the real-time coal image data and physical parameter data that have undergone preprocessing and feature extraction into the trained large-lump coal recognition model, so that the model can analyze and judge the input data based on the learned feature patterns and classification rules, identify the large lumps of coal, and mark the location and size information of the large lumps of coal; at the same time, the recognition results are transmitted to the application module for users to conduct real-time monitoring, statistical analysis and production decision-making.

[0066] The specific steps of the image preprocessing module are as follows:

[0067] Step 1: Denoising: Use median filtering and Gaussian filtering algorithms to remove salt and pepper noise and Gaussian noise in the image to improve image clarity;

[0068] Step 2: Image enhancement: Using histogram equalization and adaptive histogram equalization techniques, the image contrast is enhanced to make the edges and details of the coal blocks more prominent, facilitating subsequent feature extraction.

[0069] Step 3: Image segmentation: Use threshold segmentation, region growing, and semantic segmentation algorithms to separate the coal blocks in the coal image from the background and extract the image of a single coal block, laying the foundation for subsequent feature extraction and recognition.

[0070] The specific steps of the feature extraction module are as follows:

[0071] Step 1: Image feature extraction: For the segmented coal image, deep learning algorithms (such as convolutional neural networks) or traditional image processing methods (such as edge detection and shape description) are used to extract the coal's size (length, width, height), shape (circularity, rectangularity, etc.), and texture (roughness, directionality, etc.). At the same time, combined with the three-dimensional spatial information obtained by the lidar, the coal's volume and other three-dimensional features are calculated.

[0072] Step 2: Extraction of physical parameter features: The physical parameter data such as weight and temperature collected by the sensor are normalized and standardized to eliminate the impact of dimensional differences between different parameters. At the same time, through statistical analysis, statistical features such as mean, variance, maximum value, and minimum value are extracted to provide multi-dimensional information support for the identification of large coal lumps.

[0073] The specific steps of the model training module are as follows:

[0074] Step 1: Data labeling and partitioning: First, collect coal image data, then manually label large pieces of coal to clarify their location and size information. Then, divide the labeled data into training, validation, and test sets. The training set usually accounts for 70%-80% and is used for model parameter learning; the validation set accounts for 10%-15% and is used to adjust model hyperparameters and prevent overfitting; and the test set accounts for 10%-15% and is used to evaluate the final model performance.

[0075] Step 2: Select a suitable deep learning model: First, select a classic model in the convolutional neural network, such as ResNet, YOLO, etc., and then optimize and adjust the model structure according to actual needs to input the training set data into the model, and continuously adjust the model's weight parameters through the backpropagation algorithm and gradient descent method to minimize the error (such as the loss function value) between the model's prediction results and the labeled data.

[0076] Step 3: Model optimization and evaluation: During the training process, the model is evaluated using the validation set data. If overfitting or underfitting occurs, it will be optimized by adjusting the model structure, adding data augmentation operations, and adjusting the learning rate method. Once the model performance on the validation set is stable, the test set is used to conduct a final evaluation of the model to determine the model's generalization ability and reliability.

[0077] When used, the specific steps are as follows:

[0078] S1: Collect coal images, three-dimensional spatial information of coal, and physical parameters of coal through the data acquisition module;

[0079] S2: Establish communication connection with various devices of the equipment status information collection module and the data acquisition module to obtain key parameter information of equipment operation in real time; after establishment, the received raw data will be preprocessed by the data preprocessing module, first using the filtering algorithm to remove noise interference in the data, and then using interpolation method or statistical model to complete it according to the historical operation data of the equipment and the status information of the adjacent equipment to ensure the accuracy and completeness of the subsequent analysis data; after preprocessing, the fault threshold setting and judgment module will first set a reasonable threshold range for each monitoring parameter according to the technical parameters of the equipment and the historical data of normal operation; then compare the real-time collected and preprocessed data with the set threshold. Once the parameter is found to be out of the threshold range, the fault detection process will be triggered immediately; after judgment, the fault type location module will be used to detect abnormal parameters. It can accurately locate the fault type through fault tree analysis and neural network intelligent algorithms, combined with the changes in various equipment parameters and the fault knowledge base. After locating the fault, the fault alarm and information push module will immediately send an alarm signal after determining the fault type and location, and use the sound and light alarm device to prompt the occurrence of the fault on site. At the same time, detailed fault information will be sent to equipment maintenance personnel and relevant management personnel by SMS, email or system message push to ensure that the responsible person is notified as soon as possible. After the push, the fault repair and feedback module will be used to enable maintenance personnel to perform on-site maintenance based on the information provided after receiving the alarm information. After the repair is completed, the fault repair status can be fed back, and the equipment operating parameters can be re-collected to confirm that the equipment has returned to normal operation. At the same time, the fault knowledge base will be updated to provide richer experience data for subsequent fault diagnosis.

[0080] S3: The data collected by the data acquisition module is transmitted through the data transmission module. To ensure the stability and reliability of data transmission, data encryption and data verification technologies are used;

[0081] S4: The data collected by the data acquisition module is first received through the data receiving and diversion module, and then diverted according to the data type and subsequent processing requirements; after diversion, the collected image will be denoised, enhanced and segmented through the image preprocessing module; after processing, the segmented coal block image will be first extracted in size, shape and texture through the feature extraction module, and then the physical parameter data will be statistically extracted; after extraction, the coal image data will be first labeled and divided through the model training module, and then the appropriate deep learning model will be selected and optimized and evaluated; after that, the real-time coal image data and physical parameter data that have been preprocessed and feature extracted will be input into the trained large coal recognition model through the recognition and classification module, so that the model can analyze and judge the input data according to the learned feature patterns and classification rules, identify the large coal, and mark the position and size information of the large coal; at the same time, the recognition results will be transmitted to the application module for users to conduct real-time monitoring, statistical analysis and production decision-making.

[0082] Although the present invention has been described above with reference to embodiments, various modifications may be made thereto and equivalent components may be substituted without departing from the scope of the present invention. In particular, as long as there are no structural conflicts, the various features of the embodiments disclosed herein may be combined with each other in any manner, and the omission of an exhaustive description of such combinations in this specification is solely for the sake of space and resource conservation. Therefore, the present invention is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. A large coal intelligent identification system, characterized in that: include: A data acquisition module is used to collect coal images, three-dimensional spatial information of coal, and physical parameters of coal; The fault diagnosis module is used to monitor the operating status of each acquisition device in real time, so as to quickly locate the fault point by analyzing the device's working parameters and data transmission status; The data transmission module is used to transmit the data collected by the data acquisition module. To ensure the stability and reliability of data transmission, data encryption and data verification technologies will be used; The data processing module is used to process and analyze the collected data, including image preprocessing, feature extraction, model training, and recognition and classification; The application module is used to provide users with various application functions of the system.

2. The intelligent identification system for large coal according to claim 1, characterized in that: The fault diagnosis module includes: The equipment status information collection module is used to establish communication connections with various devices in the data acquisition module to obtain key parameter information of equipment operation in real time; The data preprocessing module is used to preprocess the received raw data. It first uses a filtering algorithm to remove noise interference in the data, and then uses interpolation or statistical models to complete the data based on the historical operation data of the device and the status information of adjacent devices to ensure the accuracy and completeness of the subsequent analysis data.

3. The intelligent identification system for large coal according to claim 2, characterized in that: The fault diagnosis module also includes: The fault threshold setting and judgment module is used to first set a reasonable threshold range for each monitoring parameter based on the technical parameters of the equipment and historical data of normal operation. It then compares the real-time collected and pre-processed data with the set threshold. If a parameter is found to exceed the threshold range, the fault detection process is immediately triggered. The fault type location module is used to accurately locate the fault type after detecting abnormal parameters through fault tree analysis, neural network intelligent algorithms, combined with the changes in various parameters of the equipment and the fault knowledge base.

4. The intelligent identification system for large coal according to claim 3, characterized in that: The fault diagnosis module also includes: The fault alarm and information push module is used to immediately send out an alarm signal after determining the fault type and location, and to notify the fault on site through the sound and light alarm device. At the same time, detailed fault information is sent to equipment maintenance personnel and relevant management personnel via SMS, email or system message push to ensure that the responsible person is notified as soon as possible. The fault repair and feedback module is used to enable maintenance personnel to perform on-site repairs based on the information provided after receiving alarm information. After the repair is completed, it can provide feedback on the fault repair status and re-collect equipment operating parameters to confirm that the equipment has returned to normal operating status. At the same time, it updates the fault knowledge base to provide richer experience data for subsequent fault diagnosis.

5. The intelligent identification system for large coal according to claim 1, characterized in that: The data processing module includes: The data receiving and diversion module is used to first receive the data collected by the data acquisition module, and then divert it according to the data type and subsequent processing requirements; Image preprocessing module, used to perform denoising, enhancement and segmentation on the collected images; The feature extraction module is used to first extract the size, shape and texture of the segmented coal block image, and then perform statistical feature extraction on the physical parameter data; The model training module is used to first label and segment the coal image data, then select the appropriate deep learning model and perform optimization and evaluation; The recognition and classification module is used to input the real-time coal image data and physical parameter data that have undergone preprocessing and feature extraction into the trained large-lump coal recognition model, so that the model can analyze and judge the input data based on the learned feature patterns and classification rules, identify the large lumps of coal, and mark the location and size information of the large lumps of coal; at the same time, the recognition results are transmitted to the application module for users to conduct real-time monitoring, statistical analysis and production decision-making.

6. The intelligent identification system for large coal according to claim 5, characterized in that: The specific steps of the image preprocessing module are as follows: Step 1: Denoising: Use median filtering and Gaussian filtering algorithms to remove salt and pepper noise and Gaussian noise in the image to improve image clarity; Step 2: Image enhancement: Using histogram equalization and adaptive histogram equalization techniques, the image contrast is enhanced to make the edges and details of the coal blocks more prominent, facilitating subsequent feature extraction. Step 3: Image segmentation: Use threshold segmentation, region growing, and semantic segmentation algorithms to separate the coal blocks in the coal image from the background and extract the image of a single coal block, laying the foundation for subsequent feature extraction and recognition.

7. The intelligent identification system for large coal according to claim 5, characterized in that: The specific steps of the feature extraction module are as follows: Step 1: Image feature extraction: For the segmented coal block images, deep learning algorithms or traditional image processing methods are used to extract the size, shape, and texture features of the coal blocks. Simultaneously, the 3D spatial information acquired by the LiDAR is combined to calculate the 3D features of the coal blocks, such as their volume. Step 2: Extraction of physical parameter features: The physical parameter data collected by the sensor is normalized and standardized to eliminate the impact of dimensional differences between different parameters. At the same time, statistical features are extracted through statistical analysis to provide multi-dimensional information support for the identification of large coal blocks.

8. The intelligent identification system for large coal according to claim 5, characterized in that: The specific steps of the model training module are as follows: Step 1: Data labeling and segmentation: First, collect coal image data, then manually label large pieces of coal to clarify their location and size information; then divide the labeled data into training, validation, and test sets; Step 2: Select an appropriate deep learning model: First, select a classic model in the convolutional neural network. Then, optimize and adjust the model structure according to actual needs. Input the training set data into the model. Use the backpropagation algorithm and gradient descent method to continuously adjust the model weight parameters to minimize the error between the model's prediction results and the labeled data. Step 3: Model optimization and evaluation: During the training process, the model is evaluated using the validation set data. If overfitting or underfitting occurs, it will be optimized by adjusting the model structure, adding data augmentation operations, and adjusting the learning rate method. Once the model performance on the validation set is stable, the test set is used to conduct a final evaluation of the model to determine the model's generalization ability and reliability.

Images