Coal exploration area identification method, system and equipment

Abstract

The invention provides a coal exploration area identification method, system and device. The method comprises the following steps: firstly, collecting a plurality of remote sensing images of a target exploration area; performing semantic segmentation on each remote sensing image to obtain contour features of different earth surface objects in each remote sensing image; further determining a semantic association relationship between different earth surface objects in each remote sensing image, and then determining the recognition contribution degree of each remote sensing image according to the spectral reflection characteristics of each remote sensing image and all semantic association relationships; extracting the area edge of coal distribution from all the thermal infrared band remote sensing images through the temperature constraint of each thermal infrared band remote sensing image; and further determining a credible exploration area of the coal distribution according to the area edge of the coal distribution and the recognition contribution degree of each remote sensing image. By adopting the scheme of the invention, the influence of coal seam mineral composition reflection characteristics on potential mining area identification can be reduced, and thus the accuracy of identifying coal exploration area distribution is improved.

Description

Technical Field

[0001] This application relates to the field of image recognition technology. More specifically, this application relates to a method, system, and device for identifying coal exploration areas. Background Art

[0002] Image recognition uses artificial intelligence technology to automatically identify and classify objects, scenes, or activities by analyzing and processing information in images. In the coal mining industry, image recognition technology is widely used in fields such as mine area monitoring, coal seam detection, and equipment maintenance. Through images captured by cameras or drones installed in the mine area and combined with deep learning and pattern recognition algorithms, the system can automatically identify changes in the mine area environment, equipment failures, the distribution of coal seams, and potential safety hazards. This not only improves the production efficiency of coal mines but also effectively reduces the safety risks of manual operations, promoting the intelligent development of coal mines.

[0003] The existing identification of coal exploration areas relies on remote sensing technology. Common technologies include optical remote sensing, infrared remote sensing, radar remote sensing, and lidar, etc. By analyzing remote sensing data, the differences between coal and the surrounding environment are identified. At the same time, thermal infrared remote sensing images can detect coal mine spontaneous combustion phenomena or surface temperature anomalies to help identify potential mining areas. However, there are often various minerals in coal seams, such as mudstone, sandstone, gypsum, etc. The reflection characteristics of these minerals may be very close to those of coal, resulting in difficulty in effectively distinguishing them in remote sensing images. Especially when using the short-wave infrared band and mid-infrared band for mineral analysis, the similarity between mineral components may affect the accuracy of the analysis, leading to an inability to accurately identify the distribution of coal exploration areas. Therefore, how to reduce the influence of the reflection characteristics of coal seam mineral components on the identification of potential mining areas, thereby improving the accuracy of identifying the distribution of coal exploration areas. Summary of the Invention

[0004] This application provides a method, system, and device for identifying coal exploration areas, which can reduce the influence of the reflection characteristics of coal seam mineral components on the identification of potential mining areas, thereby improving the accuracy of identifying the distribution of coal exploration areas.

[0005] In a first aspect, this application provides a method for identifying coal exploration areas, including the following steps: Collect multiple remote sensing images of the target exploration area in different bands; Based on a preset image semantic segmentation model, perform semantic segmentation on each remote sensing image, and then obtain the contour features of different surface objects in each remote sensing image; Determine the semantic association relationship between different surface objects in each remote sensing image according to the semantic differences and contour features between different surface objects in each remote sensing image, and determine the recognition contribution degree of each remote sensing image in the process of identifying the exploration area according to the spectral reflection characteristics of each remote sensing image under the same band and all semantic association relationships; Obtain remote sensing images in the thermal infrared band at different time periods, and extract the regional edge of coal distribution in the target exploration area from all thermal infrared band remote sensing images through the temperature constraint of each thermal infrared band remote sensing image; Determine the credible exploration area of coal distribution in the target exploration area according to the regional edge of coal distribution in the target exploration area and the recognition contribution degree of each remote sensing image in the process of identifying the exploration area.

[0006] In some embodiments, performing semantic segmentation on each remote sensing image based on a preset image semantic segmentation model, and further obtaining the contour features of different surface objects in each remote sensing image specifically includes: Preprocess all remote sensing images to obtain multiple preprocessed remote sensing images; Initialize the image semantic segmentation model; Segment each preprocessed remote sensing image according to the preset image semantic segmentation model to obtain the pixel labels of different surface objects in each preprocessed remote sensing image; Extract edge features from the pixel labels of different surface objects in each preprocessed remote sensing image to obtain the contour features of different surface objects in each remote sensing image.

[0007] In some embodiments, determining the semantic association relationship between different surface objects in each remote sensing image according to the semantic differences and contour features between different surface objects in each remote sensing image specifically includes: Select a remote sensing image as the selected remote sensing image; Determine the semantic differences between different surface objects in the selected remote sensing image; Extract the semantic relevance between each surface object in the selected remote sensing image and coal based on a large-scale pre-trained language model; Determine the semantic association relationship between different surface objects in the selected remote sensing image through the semantic relevance between each surface object in the selected remote sensing image and coal, the semantic differences between different surface objects in the remote sensing image, and the contour features of different surface objects in the selected remote sensing image; Continue to determine the semantic association relationship between different surface objects in the remaining remote sensing images.

[0008] In some embodiments, determining the recognition contribution degree of each remote sensing image in the process of identifying the exploration area according to the spectral reflection characteristics of each remote sensing image in the same band and all semantic association relationships specifically includes: Select remote sensing images in the same band from all-band remote sensing images; Determine the spectral differences between each remote sensing image in the same band according to the spectral reflection characteristics of all remote sensing images in the same band; Determine the association confidence value of each remote sensing image in the process of identifying the exploration area through the spectral differences between each remote sensing image in all the same bands and all semantic association relationships; Determine the recognition contribution degree of each remote sensing image in the process of identifying the exploration area according to the association confidence value of each remote sensing image in the process of identifying the exploration area.

[0009] In some embodiments, extracting the regional edge of coal distribution in the target exploration area from all thermal infrared band remote sensing images through the temperature constraint of each thermal infrared band remote sensing image specifically includes: Determine the temperature constraint of each remote sensing image in the thermal infrared band in the same time period according to the brightness temperature of all remote sensing images in the thermal infrared band in the same time period, and then obtain the temperature constraint of each remote sensing image in the thermal infrared band in different time periods; Extract the edge contour of the coal distribution area in each thermal infrared band remote sensing image based on a preset coal mine area edge detection model; Smooth the edge contour of the coal distribution area in each thermal infrared band remote sensing image according to the temperature constraint of each thermal infrared band remote sensing image in different time periods to obtain the edge smooth contour of the coal distribution area in each thermal infrared band remote sensing image; Fuse the edge smooth contours of the coal distribution areas in all thermal infrared band remote sensing images to obtain the regional edge of coal distribution in the target exploration area.

[0010] In some embodiments, determining the credible exploration area of coal distribution in the target exploration area according to the regional edge of coal distribution in the target exploration area and the recognition contribution degree of each remote sensing image in the process of identifying the exploration area specifically includes: Perform linear fitting on the recognition contribution degrees of all remote sensing images in the process of identifying the exploration area to obtain a fitting curve of the recognition contribution degree; Determine the credible exploration area of coal distribution in the target exploration area according to the fitting curve of the recognition contribution degree and the regional edge of coal distribution in the target exploration area.

[0011] In some embodiments, the bands include visible light band, near infrared band, short wave infrared band, thermal infrared band and middle wave infrared band.

[0012] In a second aspect, the present application provides a coal exploration area identification system, including: An acquisition module for acquiring multiple remote sensing images of a target exploration area in different bands; A processing module for performing semantic segmentation on each remote sensing image based on a preset image semantic segmentation model, and then obtaining the contour features of different surface objects in each remote sensing image; The processing module is further configured to determine the semantic association relationship between different surface objects in each remote sensing image according to the semantic differences between different surface objects in each remote sensing image and the contour features of different surface objects, and determine the recognition contribution degree of each remote sensing image in the process of exploring area identification according to the spectral reflection features of each remote sensing image in the same band and all the semantic association relationships; The processing module is further configured to obtain remote sensing images in the thermal infrared band at different time periods, and extract the regional edge of coal distribution in the target exploration area from all the remote sensing images in the thermal infrared band through the temperature constraint of each remote sensing image in the thermal infrared band; An execution module for determining a credible exploration area of coal distribution in the target exploration area according to the regional edge of coal distribution in the target exploration area and the recognition contribution degree of each remote sensing image in the process of exploring area identification.

[0013] In a third aspect, the present application provides a computer device, which includes a memory and a processor. The memory stores code, and the processor is configured to obtain the code and execute the above-mentioned coal exploration area identification method.

[0014] In a fourth aspect, the present application provides a computer-readable storage medium, which stores a computer program, and when the computer program is executed by a processor, the above-mentioned coal exploration area identification method is implemented.

[0015] The technical solutions provided by the disclosed embodiments of the present application have the following beneficial effects: In the coal exploration area identification method, system and equipment provided by the present application, first, multiple remote sensing images of the target exploration area in different bands are collected; based on a preset image semantic segmentation model, semantic segmentation is performed on each remote sensing image, and then the contour features of different surface objects in each remote sensing image are obtained; according to the semantic differences between different surface objects in each remote sensing image and the contour features of different surface objects, the semantic association relationship between different surface objects in each remote sensing image is determined, and according to the spectral reflection characteristics of each remote sensing image in the same band and all semantic association relationships, the recognition contribution degree of each remote sensing image in the process of exploration area identification is determined; remote sensing images in the thermal infrared band at different time periods are obtained, and the regional edge of coal distribution in the target exploration area is extracted from all thermal infrared band remote sensing images through the temperature constraint of each thermal infrared band remote sensing image; according to the regional edge of coal distribution in the target exploration area and the recognition contribution degree of each remote sensing image in the process of exploration area identification, a credible exploration area of coal distribution in the target exploration area is determined.

[0016] It can be seen that in this application, the credible exploration area of coal distribution in the target exploration area can be determined according to the regional edge of coal distribution in the target exploration area and the recognition contribution degree of each remote sensing image in the process of exploration area recognition. Firstly, multiple remote sensing images of different bands are collected. The collection of images with different bands not only provides different reflection characteristic information for ground object classification. Secondly, each remote sensing image is processed using a preset image semantic segmentation model to extract the contour features of different surface objects, and the semantic association relationship is analyzed based on the semantic differences of surface objects. Through this step, the reflection characteristics of coal and other ground objects can be effectively distinguished from the images, eliminating the interference of mineral components on regional recognition. Semantic segmentation not only improves the recognition accuracy but also helps to avoid misjudgment caused by similar reflection characteristics of ore layers through the distinction of different ground objects, ensuring that the coal area can be accurately recognized. Then, on this basis, combining the spectral reflection characteristics and semantic association relationship of each remote sensing image, the recognition contribution degree of each image in the process of coal exploration area recognition is calculated. This quantitative analysis method helps to measure the importance of different bands for coal area recognition, thereby reasonably weighting the contributions of each remote sensing image and further optimizing the precise positioning of the coal area. Through the comprehensive analysis of multi-band images, the recognition deviation that may be caused by differences in mineral reflection characteristics is reduced. In addition, by obtaining remote sensing images in the thermal infrared band at different time periods and using temperature constraints to extract the edge of the target exploration area, the boundary recognition accuracy of the coal area is further improved. The application of remote sensing images in the thermal infrared band helps to clearly distinguish the coal area from surrounding ground objects. Especially at night or when the temperature difference is large, the thermal characteristics of the coal area can be more effectively recognized, thus effectively enhancing the precision of the coal area edge and reducing the interference of thermal reflection errors on the recognition result. Finally, by comprehensively analyzing the regional edge of the target exploration area and the recognition contribution degree of each image, the recognition credibility of the coal area is obtained. This comprehensive evaluation process ensures that the positioning result of the coal area has a high credibility, reduces the recognition error that may be brought by the reflection characteristics of coal seam mineral components, and improves the accuracy of the distribution of the coal exploration area. In summary, the solution of this application can reduce the influence of the reflection characteristics of coal seam mineral components on the recognition of potential mining areas, thereby improving the accuracy of recognizing the distribution of coal exploration areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 is an exemplary flowchart of a method for identifying a coal exploration area according to some embodiments of the present application; Figure 2 is a schematic flowchart of a process for determining contour features according to some embodiments of the present application; Figure 3 is a schematic flowchart of a process for determining the recognition contribution degree according to some embodiments of the present application; Figure 4It is a schematic structural diagram of a coal exploration area identification system shown in some embodiments of the present application; Figure 5 It is a schematic structural diagram of a computer device for implementing a coal exploration area identification method shown in some embodiments of the present application. Detailed implementation manners

[0018] To better understand the technical solution of the present application, the technical solution of the present application will be described in detail below in conjunction with the accompanying drawings of the specification and specific implementation manners.

[0019] Refer to Figure 1 , this figure is an exemplary flowchart of a coal exploration area identification method shown in some embodiments of the present application. The coal exploration area identification method 100 mainly includes the following steps: In step 101, multiple remote sensing images of the target exploration area in different bands are collected.

[0020] Specifically, multiple remote sensing images of the target exploration area in different bands are collected by remote sensing equipment carried on an unmanned aerial vehicle.

[0021] It should be noted that the bands described in the present application include visible light band, near-infrared band, short-wave infrared band, thermal infrared band, and medium-wave infrared band.

[0022] In step 102, semantic segmentation is performed on each remote sensing image based on a preset image semantic segmentation model, and then the contour features of different surface objects in each remote sensing image are obtained.

[0023] In some embodiments, refer to Figure 2 shown, this figure is a schematic flowchart for determining contour features in some embodiments of the present application. In this embodiment, semantic segmentation is performed on each remote sensing image based on a preset image semantic segmentation model, and then the contour features of different surface objects in each remote sensing image can be implemented by the following steps: First, in step 1021, all remote sensing images are preprocessed to obtain multiple preprocessed remote sensing images; Secondly, in step 1022, the image semantic segmentation model is initialized; Then, in step 1023, each preprocessed remote sensing image is segmented according to the preset image semantic segmentation model to obtain pixel labels of different surface objects in each preprocessed remote sensing image; Finally, in step 1024, edge features of the pixel labels of different surface objects in each preprocessed remote sensing image are extracted to obtain the contour features of different surface objects in each remote sensing image.

[0024] It should be noted that in this application, the image semantic segmentation model can adopt a semantic segmentation model (such as FCN, U-Net, DeepLabV3+, etc., which are not limited here). Training data is obtained by annotating different surface objects in historical remote sensing images. Then, the semantic segmentation model after training is used as the image semantic segmentation model to complete the initialization of the image semantic segmentation model. Among them, the surface objects include different types such as mountain fold areas, mountain fault areas, sedimentary basins, mountain fault lines, rivers, lakes, wetlands, vegetation, ore rock layers, soil, coal mine waste residues, coal blocks, thermal infrared anomaly areas, and coal distribution areas.

[0025] When specifically implemented, first, preprocess all remote sensing images. The preprocessing includes denoising (such as Gaussian blur denoising), normalization, contrast enhancement (such as contrast stretching), and cropping and scaling. And all the remote sensing images obtained after preprocessing are used as preprocessed remote sensing images. Secondly, input each preprocessed remote sensing image into a preset image semantic segmentation model. Through the image semantic segmentation model, semantic segmentation is performed on each preprocessed remote sensing image to obtain different surface objects in each preprocessed remote sensing image. Then, the pixel labels of different surface objects in each preprocessed remote sensing image are output through the image semantic segmentation model, where the pixel label represents a vector composed of the categories to which the pixel points belong. Then, edge extraction is performed on the pixel labels of different surface objects in each preprocessed remote sensing image through an edge detection algorithm (such as Canny edge detection) to obtain the edge contours of different surface objects in each preprocessed remote sensing image. Further, Hu moments are calculated for the edge contours of different surface objects in each preprocessed remote sensing image, and the calculated values are used as the contour features of different surface objects in each preprocessed remote sensing image. In other embodiments, other methods can also be used to implement this, which will not be elaborated here.

[0026] It should be noted that in this application, the contour feature represents the geometric feature of the boundary of the surface object in the remote sensing image.

[0027] In step 103, according to the semantic differences between different surface objects in each remote sensing image and the contour features of different surface objects, determine the semantic association relationship between different surface objects in each remote sensing image. According to the spectral reflection features of each remote sensing image in the same band and all the semantic association relationships, determine the recognition contribution degree of each remote sensing image in the process of identifying the exploration area.

[0028] In some embodiments, the determination of the semantic association relationship between different surface objects in each remote sensing image according to the semantic differences between different surface objects in each remote sensing image and the contour features of different surface objects can be implemented by the following steps: Select a remote sensing image as the selected remote sensing image; Determine the semantic differences between different surface objects within the selected remote sensing image; Extract the semantic relevance between each surface object within the selected remote sensing image and coal based on a large-scale pre-trained language model; Determine the semantic association relationship between different surface objects within the selected remote sensing image based on the semantic relevance between each surface object within the selected remote sensing image and coal, the semantic differences between different surface objects within the remote sensing image, and the contour features of different surface objects within the selected remote sensing image; Continue to determine the semantic association relationship between different surface objects within the remaining remote sensing images.

[0029] It should be noted that in this application, the large-scale pre-trained language model used is the GPT-4 language model. Additionally, it should be noted that the semantic difference in this application represents the parameter that shows the difference in semantics of surface objects with similar names within the remote sensing image.

[0030] In specific implementation, first, use the word embedding technique (such as Word2Vec) to convert the names of different surface objects within the selected remote sensing image into vectors, and thus regard the obtained vectors as the name vectors of different surface objects within the selected remote sensing image. Then, calculate the Euclidean distance between the name vectors of every two surface objects within the selected remote sensing image, and take the smallest Euclidean distance among all Euclidean distances as the semantic difference between different surface objects within the selected remote sensing image. Secondly, input the name of coal into the large-scale pre-trained language model, extract the keywords that are closest to the name of coal through the large-scale pre-trained language model, convert the keywords into keyword vectors through the word embedding technique, further calculate the cosine similarity between the keyword vectors and the name vectors of each surface object within the selected remote sensing image, and regard the obtained cosine similarities as the semantic relevance between each surface object within the selected remote sensing image and coal. Then, multiply the contour feature of each surface object within the selected remote sensing image by the semantic relevance between each surface object within the selected remote sensing image and coal, sum all the multiplied values, further divide the summed value by the semantic difference between different surface objects within the selected remote sensing image, and take the divided value as the semantic association relationship between different surface objects within the selected remote sensing image.

[0031] It should be noted that the semantic relevance in this application represents the association strength between the surface object within the remote sensing image and coal at the semantic level. Additionally, the semantic association relationship represents the feature that the semantics of each surface object within the remote sensing image are interrelated during the process of identifying the coal exploration area. Through the semantic association relationship, it is possible to identify which surface objects have a strong semantic relevance to the coal exploration area, thereby reducing the interference of irrelevant ground objects (such as limestone, desert, etc.) and avoiding misidentifying them as potential coal mining areas.

[0032] In some embodiments, with reference to Figure 3 As shown, the figure is a schematic flowchart for determining the recognition contribution degree in some embodiments of the present application. In this embodiment, the recognition contribution degree of each remote sensing image in the exploration area recognition process can be determined according to the spectral reflection characteristics of each remote sensing image in the same band and all semantic association relationships, and can be implemented by the following steps: Select remote sensing images in the same band from all-band remote sensing images; Determine the spectral differences between each remote sensing image in the same band according to the spectral reflection characteristics of all remote sensing images in the same band; Determine the association confidence value of each remote sensing image in the exploration area recognition process through the spectral differences between each remote sensing image in the same band and all semantic association relationships; Determine the recognition contribution degree of each remote sensing image in the exploration area recognition process according to the association confidence value of each remote sensing image in the exploration area recognition process.

[0033] It should be noted that the spectral reflection characteristics described in the present application represent the distribution characteristics of the electromagnetic wave reflection of the surface objects in the remote sensing image in a specific band, and the spectral reflection characteristics of the remote sensing image in the present application can be described by the reflectivity of a specific band.

[0034] Specifically, in implementation, first, perform differential processing on the spectral reflection characteristics of all remote sensing images in the same band, then sum all the differential values, and use the sum value as the spectral difference between each remote sensing image in the same band. The spectral difference represents the difference degree between the spectral reflection characteristics of each remote sensing image in the same band; secondly, calculate the negative exponential function with the natural logarithm e as the base for the spectral difference between each remote sensing image in the same band, and then multiply the obtained values after calculating the negative exponential function with the natural logarithm e as the base by the semantic association relationships between different surface objects in each remote sensing image in the same band, and use the multiplied values as the association confidence value of each remote sensing image in the exploration area recognition process in the same band. Furthermore, perform the same processing on each remote sensing image in all bands, so as to obtain the association confidence value of each remote sensing image in the exploration area recognition process. Among them, the association confidence value represents an index for evaluating the credibility of the association between the remote sensing image and the coal distribution; then, divide the association confidence value of each remote sensing image in the exploration area recognition process by the sum of the association confidence values of all remote sensing images in the exploration area recognition process, and use the obtained division values as the recognition contribution degree of each remote sensing image in the exploration area recognition process. In other embodiments, other methods can also be used for implementation, which will not be elaborated here.

[0035] It should be noted that the recognition contribution degree described in this application represents the influence degree of the remote sensing image on the recognition result of the coal distribution area during the coal exploration area recognition process. The greater the recognition contribution degree, the higher the influence degree of the remote sensing image on the recognition result of the coal distribution area during the coal exploration area recognition process, and vice versa. Through the recognition contribution degree, the interference brought by the reflection characteristics of mineral components to the exploration result can be reduced, thereby improving the recognition accuracy of the coal area.

[0036] In step 104, remote sensing images in the thermal infrared band at different time periods are obtained, and the regional edge of coal distribution within the target exploration area is extracted from all the remote sensing images in the thermal infrared band through the temperature constraint of each remote sensing image in the thermal infrared band.

[0037] Specifically, multiple remote sensing images in the thermal infrared band are collected during the day (10:00 - 16:00) and at night (18:00 - 02:00) in summer and winter respectively. It should be noted that in this application, remote sensing images need to be collected at different time points at the same location.

[0038] In some embodiments, the extraction of the regional edge of coal distribution within the target exploration area from all the remote sensing images in the thermal infrared band through the temperature constraint of each remote sensing image in the thermal infrared band can be implemented by the following steps: Determine the temperature constraint of each remote sensing image in the thermal infrared band within the same time period according to the brightness temperature of all remote sensing images in the thermal infrared band within the same time period, and then obtain the temperature constraint of each remote sensing image in the thermal infrared band at different time periods; Based on a preset coal mine area edge detection model, extract the edge contour of the coal distribution area within each remote sensing image in the thermal infrared band; Smooth the edge contour of the coal distribution area within each remote sensing image in the thermal infrared band according to the temperature constraint of each remote sensing image in the thermal infrared band at different time periods to obtain the edge smooth contour of the coal distribution area within each remote sensing image in the thermal infrared band; Fuse the edge smooth contours of the coal distribution areas within all remote sensing images in the thermal infrared band to obtain the regional edge of coal distribution within the target exploration area.

[0039] It should be noted that the temperature constraint described in this application represents the parameter value that restricts the temperature change of the coal distribution area within the remote sensing image in the thermal infrared band. In addition, it should also be noted that the coal mine area edge detection model adopted in this application is a U-Net deep learning model, which can be trained through multiple infrared band remote sensing images marked with coal exploration areas. Among them, the coal mine area edge detection model can output the edge contour of the coal distribution area within each infrared band remote sensing image.

[0040] In specific implementation, first, the radiance of all remote sensing images in the thermal infrared band during the same period is converted into brightness temperature using a radiative transfer model (such as Moderate Resolution Atmospheric Transmission, MODTRAN), so as to obtain the brightness temperature of all remote sensing images in the thermal infrared band during the same period. Among them, the pixel value of the thermal infrared band remote sensing image represents the radiance of the ground surface. The maximum brightness temperature and the minimum brightness temperature are selected from the brightness temperatures of all remote sensing images in the thermal infrared band during the same period, and the difference between the maximum brightness temperature and the minimum brightness temperature is used as the temperature constraint for each remote sensing image in the thermal infrared band during the same period, thereby obtaining the temperature constraints for each remote sensing image in the thermal infrared band during different periods. Secondly, each thermal infrared band remote sensing image is input into a preset coal mine area edge detection model, and the output result of the coal mine area edge detection model is used as the edge contour of the coal distribution area in each thermal infrared band remote sensing image. Then, the Gaussian filtering algorithm is used to smooth the edge contour of the coal distribution area in each thermal infrared band remote sensing image, so as to obtain the edge smoothed contour of the coal distribution area in each thermal infrared band remote sensing image. Among them, when processing each thermal infrared band remote sensing image, the standard deviation of the Gaussian filtering algorithm is set to the quotient between the average brightness of the thermal infrared band remote sensing image and the temperature constraint of the thermal infrared band remote sensing image. Finally, OpenCV is used to perform a logical AND operation on the edge smoothed contours of the coal distribution areas in all thermal infrared band remote sensing images, and the result obtained after the logical AND operation is used as the regional edge of the coal distribution in the target exploration area. In other embodiments, other methods can also be used to achieve this, which will not be elaborated here.

[0041] It should be noted that the edge smoothed contour described in this application represents the contour obtained after the edge contour of the coal distribution area in the thermal infrared band remote sensing image is smoothed. In addition, it should also be noted that the regional edge described in this application represents the edge distribution of the coal area in the image of the target exploration area.

[0042] In step 105, the reliable exploration area of the coal distribution in the target exploration area is determined according to the regional edge of the coal distribution in the target exploration area and the recognition contribution degree of each remote sensing image during the exploration area recognition process.

[0043] In some embodiments, determining the reliable exploration area of the coal distribution in the target exploration area according to the regional edge of the coal distribution in the target exploration area and the recognition contribution degree of each remote sensing image during the exploration area recognition process can be achieved by the following steps: Perform linear fitting on the recognition contribution degrees of all remote sensing images during the exploration area recognition process to obtain the fitting curve of the recognition contribution degree; Determine the credible exploration area of coal distribution in the target exploration area according to the fitting curve for identifying the contribution degree and the regional edge of coal distribution in the target exploration area.

[0044] In specific implementation, first, use an existing linear fitting algorithm (such as the least squares support vector machine algorithm) to linearly fit the identification contribution degrees of all remote sensing images during the exploration area identification process, and use the obtained fitting curve as the fitting curve for the identification contribution degree. Among them, each value on the fitting curve is used as the fitting value of the identification contribution degree, and each fitting value of the identification contribution degree corresponds to an identification contribution degree, and each identification contribution degree corresponds to a remote sensing image. Secondly, screen out all remote sensing images in the thermal infrared band from the remote sensing images of different bands obtained in step 101, and use all remote sensing images in the thermal infrared band as contributing remote sensing images. Screen out the fitting values of the identification contribution degrees corresponding to each contributing remote sensing image from the fitting curve of the identification contribution degree (where each contributing remote sensing image corresponds to an identification contribution degree, and each identification contribution degree corresponds to a fitting value of the identification contribution degree, that is, each contributing remote sensing image corresponds to a fitting value of the identification contribution degree on the fitting curve). For each screened fitting value of the identification contribution degree, subtract the identification contribution degree corresponding to the fitting value of the identification contribution degree, then take the absolute value of the subtracted value, and use the obtained absolute value as the edge adjustment factor, thereby obtaining multiple edge adjustment factors. Further sum all the edge adjustment factors, and use the obtained sum value as the scaling factor for the regional edge of coal distribution in the target exploration area. Then use OpenCV to perform scaling processing on the regional edge of coal distribution in the target exploration area, and use the obtained area after scaling processing as the credible exploration area of coal distribution in the target exploration area.

[0045] It should be noted that the credible exploration area in this application represents the area with a relatively high credibility of coal distribution in the target exploration area.

[0046] In addition, on the other hand of this application, in some embodiments, this application provides a coal exploration area identification system. Refer to Figure 4 , this figure is a schematic structural diagram of a coal exploration area identification system according to some embodiments of this application. The coal exploration area identification system 400 includes: an acquisition module 401, a processing module 402, and an execution module 403, which are described as follows: The acquisition module 401. In this application, the acquisition module 401 is mainly used to acquire multiple remote sensing images of the target exploration area in different bands. The processing module 402. In this application, the processing module 402 is used to perform semantic segmentation on each remote sensing image based on a preset image semantic segmentation model, and then obtain the contour features of different surface objects in each remote sensing image. It should be noted that the processing module 402 in the present application is further configured to determine the semantic association relationship between different surface objects in each remote sensing image according to the semantic differences and contour features between different surface objects in each remote sensing image, and determine the recognition contribution degree of each remote sensing image in the process of identifying the exploration area according to the spectral reflection characteristics of each remote sensing image in the same band and all the semantic association relationships; In addition, the processing module 402 in the present application is further configured to obtain remote sensing images in the thermal infrared band at different time periods, and extract the regional edge of coal distribution in the target exploration area from all the remote sensing images in the thermal infrared band through the temperature constraint of each remote sensing image in the thermal infrared band; Execution module 403. In the present application, the execution module 403 is mainly configured to determine the reliable exploration area of coal distribution in the target exploration area according to the regional edge of coal distribution in the target exploration area and the recognition contribution degree of each remote sensing image in the process of identifying the exploration area.

[0047] In addition, the present application also provides a computer device, which includes a memory and a processor. The memory stores code, and the processor is configured to obtain the code and execute the above-mentioned coal exploration area recognition method.

[0048] In some embodiments, refer to Figure 5 This figure is a schematic structural diagram of a computer device for implementing the coal exploration area recognition method according to some embodiments of the present application. The coal exploration area recognition method in the above embodiments can be implemented by Figure 5 The computer device shown. The computer device 500 includes at least one processor 501, a communication bus 502, a memory 503, and at least one communication interface 504.

[0049] The processor 501 may be a general-purpose central processing unit (CPU), or an application-specific integrated circuit (ASIC), or one or more for controlling the execution of the coal exploration area recognition method in the present application.

[0050] The communication bus 502 can be used to transmit information between the above components.

[0051] The memory 503 can be a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM), or other types of dynamic storage devices that can store information and instructions. It can also be an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM), or other optical disc storage, optical disc storage (including compact discs, laser discs, optical discs, digital versatile discs, Blu-ray discs, etc.), a magnetic disk, or other magnetic storage devices, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory 503 can exist independently and be connected to the processor 501 through the communication bus 502. The memory 503 can also be integrated with the processor 501.

[0052] Among them, the memory 503 is used to store the program code for executing the solution of this application and is controlled by the processor 501 for execution. The processor 501 is used to execute the program code stored in the memory 503. The program code can include one or more software modules. The methods described in the above method embodiments can be implemented by one or more software modules in the program code in the processor 501 and the memory 503.

[0053] The communication interface 504 uses any device such as a transceiver to communicate with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area networks (WLAN), etc.

[0054] In a specific implementation, as an embodiment, the computer device can include multiple processors, and each of these processors can be a single-core (single-CPU) processor or a multi-core (multi-CPU) processor. Here, the processor can refer to one or more devices, circuits, and / or processing cores for processing data (such as computer program instructions).

[0055] The computer device described above can be a general-purpose computer device or a special-purpose computer device. In specific implementations, the computer device can be a desktop computer, a laptop computer, a network server, a personal digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device, or an embedded device. The embodiments of the present application do not limit the type of the computer device.

[0056] In addition, the present application also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the above-mentioned coal exploration area recognition method is implemented.

[0057] Although the preferred embodiments of the present application have been described, those skilled in the art can make additional changes and modifications once they know the basic creative concept. Therefore, the appended claims are intended to be construed to include the preferred embodiments as well as all changes and modifications falling within the scope of the present application.

[0058] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalent technologies, the present application is also intended to include these modifications and variations.

Claims

1. A method for identifying a coal exploration area, characterized in that: The steps include: Collect multiple remote sensing images of the target exploration area in different bands; Perform semantic segmentation on each remote sensing image based on a preset image semantic segmentation model, and then obtain the contour features of different surface objects in each remote sensing image; Determine the semantic association relationship between different surface objects in each remote sensing image according to the semantic differences between different surface objects in each remote sensing image and the contour features of different surface objects, and determine the recognition contribution of each remote sensing image in the exploration area recognition process according to the spectral reflectance characteristics of each remote sensing image in the same band and all the semantic association relationships; Acquire thermal infrared band remote sensing images in different time periods, and extract the regional edge of coal distribution in the target exploration area from all thermal infrared band remote sensing images through the temperature constraint of each thermal infrared band remote sensing image; The credible exploration area of ​​coal distribution in the target exploration area is determined according to the regional edge of coal distribution in the target exploration area and the recognition contribution of each remote sensing image in the exploration area recognition process.

2. The method according to claim 1, characterized in that Based on the preset image semantic segmentation model, each remote sensing image is semantically segmented, and the contour features of different surface objects in each remote sensing image are obtained, including: Preprocessing all remote sensing images to obtain multiple preprocessed remote sensing images; Initialize the image semantic segmentation model; Segment each preprocessed remote sensing image according to a preset image semantic segmentation model to obtain pixel labels of different surface objects in each preprocessed remote sensing image; Edge features are extracted from pixel labels of different surface objects in each preprocessed remote sensing image to obtain contour features of different surface objects in each remote sensing image.

3. The method according to claim 1, characterized in that Determining the semantic association relationship between different surface objects in each remote sensing image according to the semantic differences between different surface objects in each remote sensing image and the contour features of different surface objects specifically includes: Selecting a remote sensing image as a selected remote sensing image; Determine semantic differences between different surface objects within selected remote sensing images; Extract the semantic relevance between each surface object and coal in the selected remote sensing image based on a large-scale pre-trained language model; Determine the semantic association relationship between different surface objects in the selected remote sensing image by selecting the semantic relevance between each surface object and coal in the remote sensing image, the semantic difference between different surface objects in the remote sensing image, and the contour features of different surface objects in the selected remote sensing image; Continue to determine the semantic associations between different surface objects in the remaining remote sensing images.

4. The method according to claim 1, characterized in that According to the spectral reflectance characteristics of each remote sensing image in the same band and all semantic associations, the recognition contribution of each remote sensing image in the exploration area recognition process is determined, including: Select remote sensing images in the same band from remote sensing images in all bands; Determine the spectral difference between remote sensing images in the same band according to the spectral reflectance characteristics of all remote sensing images in the same band; The correlation confidence value of each remote sensing image in the process of exploration area identification is determined through the spectral differences and all semantic correlation relationships between the remote sensing images in all the same bands; The recognition contribution of each remote sensing image in the exploration area recognition process is determined according to the associated confidence value of each remote sensing image in the exploration area recognition process.

5. The method according to claim 1, characterized in that The regional edge of coal distribution in the target exploration area is extracted from all thermal infrared band remote sensing images through the temperature constraint of each thermal infrared band remote sensing image, specifically including: According to the brightness temperature of all remote sensing images in the thermal infrared band in the same time period, the temperature constraint of each remote sensing image in the thermal infrared band in the same time period is determined, and then the temperature constraint of each remote sensing image in the thermal infrared band in different time periods is obtained; Based on the preset coal mine area edge detection model, the edge contour of the coal distribution area in each thermal infrared band remote sensing image is extracted; Smoothing the edge contour of the coal distribution area in each thermal infrared band remote sensing image according to the temperature constraint of each remote sensing image in the thermal infrared band in different time periods, so as to obtain the edge smooth contour of the coal distribution area in each thermal infrared band remote sensing image; The edge smoothing contours of the coal distribution areas in all thermal infrared band remote sensing images are fused to obtain the regional edge of the coal distribution in the target exploration area.

6. The method according to claim 1, characterized in that Determining the credible exploration area of ​​coal distribution in the target exploration area according to the regional edge of coal distribution in the target exploration area and the recognition contribution of each remote sensing image in the exploration area recognition process specifically includes: Perform linear fitting on the recognition contribution of all remote sensing images in the exploration area recognition process to obtain the fitting curve of recognition contribution; A credible exploration area of ​​coal distribution in the target exploration area is determined according to the fitting curve of the recognition contribution and the regional edge of the coal distribution in the target exploration area.

7. The method according to claim 1, characterized in that The wavebands include visible light waveband, near infrared waveband, short-wave infrared waveband, thermal infrared waveband and medium-wave infrared waveband.

8. A coal exploration area identification system, characterized in that: include: An acquisition module is used to acquire multiple remote sensing images of the target exploration area in different bands; A processing module, used to perform semantic segmentation on each remote sensing image based on a preset image semantic segmentation model, and then obtain contour features of different surface objects in each remote sensing image; The processing module is further used to determine the semantic association relationship between different surface objects in each remote sensing image according to the semantic differences between different surface objects in each remote sensing image and the contour features of different surface objects, and determine the recognition contribution of each remote sensing image in the exploration area recognition process according to the spectral reflectance characteristics of each remote sensing image in the same band and all the semantic association relationships; The processing module is further used to obtain thermal infrared band remote sensing images in different time periods, and extract the regional edge of coal distribution in the target exploration area from all thermal infrared band remote sensing images through the temperature constraint of each thermal infrared band remote sensing image; The execution module is used to determine the credible exploration area of ​​coal distribution in the target exploration area according to the regional edge of coal distribution in the target exploration area and the recognition contribution of each remote sensing image in the exploration area recognition process.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the method for identifying a coal exploration area according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method for identifying a coal exploration area according to any one of claims 1 to 7 is implemented.

Images