Coal seam recognition method and device and storage medium

Abstract

The invention relates to the technical field of coal seam detection, and discloses a coal seam recognition method and device and a storage medium, and the method comprises the following steps: extracting local features of a fault in a coal seam profile image through a trained fault image encoder; extracting local features and global features of the coal seam in the coal seam profile image through a trained coal seam image encoder; calculating the weights of the features of each layer of fault and coal seam through an attention mechanism, and weighting the extracted features of the fault and coal seam; fusing the features of the coal seam and the features of the fault, and converting the fused features into final features for identification; and classifying objects of the coal seam profile image according to the final features, and marking specific positions and boundaries of the coal seam and the fault. According to the method, the features of the coal seam and the fault in the coal seam profile image are fused, the recognition accuracy of the boundary of the coal seam is improved, and therefore the accuracy and comprehensiveness of coal seam position description are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of coal seam detection, and in particular to a coal seam identification method, device and storage medium. Background Art

[0002] Coal seam detection technology plays a vital role in coal mine safety production. In recent years, with the development of artificial intelligence technology, it has provided more options for the detection of industrial resources such as coal mines. Artificial intelligence technology has brought great stability and convenience to the detection of coal seams. The existing coal seam detection method based on artificial intelligence technology usually uses an artificial intelligence model trained by a coal seam profile image with the coal seam position marked to extract the coal seam features in the target coal seam profile image, and identify the location of the coal seam through the extracted features. However, since this method only extracts the features of the coal seam, the recognition accuracy of the coal seam boundary is not high, and professionals are required to make a lot of adjustments to the detection parameters, which increases the complexity of the operation and reduces the detection efficiency. Summary of the invention

[0003] In order to overcome the shortcomings of the prior art, the purpose of the present invention is to provide a coal seam identification method that can comprehensively consider the coal seam information and fault information in the coal seam profile image, improve the accuracy of coal seam boundary positioning, and improve the accuracy of coal seam position description.

[0004] To solve the above problems, the technical solution adopted by the present invention is as follows: a coal seam identification method, comprising the following steps:

[0005] Extracting local features of faults in coal seam profile images through the trained fault image encoder;

[0006] Extracting local features and global features of the coal seam in the coal seam profile image through the trained coal seam image encoder;

[0007] The weights of the features of each fault and coal seam are calculated through the attention mechanism, and the extracted features of the fault and coal seam are weighted;

[0008] The coal seam features and the fault features are fused, and the fused features are converted into final features for identification;

[0009] The objects in the coal seam profile image are classified according to the final features, and the specific locations and boundaries of the coal seams and faults are identified.

[0010] Compared with the prior art, the beneficial effect of the present invention lies in: by extracting the characteristics of the fault and the coal seam in the coal seam profile image, and then fusing the extracted coal seam characteristics and fault characteristics, combining the positional relationship between the coal seam and the fault, and learning the position information of the entire profile image, the accuracy of identifying the boundary of the coal seam is improved, and the accuracy and comprehensiveness of the description of the coal seam position are improved.

[0011] In the above-mentioned coal seam identification method, in the step of extracting the local features of the fault in the coal seam profile image by using the trained fault image encoder, the local features of the fault are extracted by using an image encoder based on a convolutional neural network and a Focal-Transformer.

[0012] In the above-mentioned coal seam identification method, in the step of extracting the local features and global features of the coal seam in the coal seam profile image by using a trained coal seam image encoder, the local features and global features of the coal seam are extracted by using an image encoder based on the ViTDet model.

[0013] In the above-mentioned coal seam identification method, the weights of the features of each fault and coal seam are calculated through the attention mechanism, and in the step of weighting the extracted features of the fault and coal seam, the features of the fault are used as the query of the attention mechanism, and the features of the coal seam are used as the key and value of the attention mechanism to calculate the weights of the features of each fault and coal seam.

[0014] In the above-mentioned coal seam identification method, in the step of fusing the coal seam features and the fault features and converting the fused features into final features for identification, the coal seam features and the fault features are fused through a Transformer-based feature decoder.

[0015] A storage medium stores a computer program, which implements the above-mentioned coal seam identification method when called and executed by a processor.

[0016] A coal seam identification device comprises a processor and a memory, wherein the processor is electrically connected to the memory, and the processor can implement the above-mentioned coal seam identification method by calling and executing a computer program in the memory.

[0017] A coal seam identification device comprises: a fault image decoder, used for extracting local features of faults in a coal seam profile image; a coal seam image decoder, used for extracting local features and global features of the coal seam in the coal seam profile image; an attention mechanism module, used for calculating the weights of features of each layer of faults and coal seams through the attention mechanism, and weighting the extracted features of the faults and coal seams; a feature fusion module, used for fusing the features of the coal seams and the features of the faults, and converting the fused features into final features for identification; and a classification module, used for classifying objects in the coal seam profile image according to the final features, and identifying the specific positions and boundaries of the coal seams and faults.

[0018] In the above-mentioned coal seam identification device, the tomographic image encoder is an image encoder based on a convolutional neural network and a Focal-Transformer, the coal seam image encoder is an image encoder based on a ViTDet model, and the feature fusion module is a feature decoder based on a Transformer.

[0019] In the above-mentioned coal seam identification device, the attention mechanism module calculates the weights of the characteristics of each fault and coal seam by using the characteristics of the fault as the query of the attention mechanism and the characteristics of the coal seam as the key and value of the attention mechanism.

[0020] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 A flow chart of a coal seam identification method according to an embodiment of the present invention;

[0022] Figure 2 4 is a schematic diagram of the structure of a coal seam identification device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0023] The embodiments of the present invention are described in detail below. Figure 1 , an embodiment of the present invention provides a coal seam identification method, comprising the following steps:

[0024] Extracting local features of faults in coal seam profile images through the trained fault image encoder;

[0025] Extracting local features and global features of the coal seam in the coal seam profile image through the trained coal seam image encoder;

[0026] The weights of the features of each fault and coal seam are calculated through the attention mechanism, and the extracted features of the fault and coal seam are weighted;

[0027] The coal seam features and the fault features are fused, and the fused features are converted into final features for identification;

[0028] The objects in the coal seam profile image are classified according to the final features, and the specific locations and boundaries of the coal seams and faults are identified.

[0029] The coal seam identification method of the embodiment of the present invention extracts fault features and coal seam features in the coal seam profile image through artificial intelligence, and fuses the fault features and coal seam features to identify the position and boundary of the coal seam in the coal seam profile image with the final fused features. Since the position information of the entire profile image is learned, the position relationship between the coal seam and the fault is combined to more accurately identify the boundary of the coal seam, thereby improving the accuracy and comprehensiveness of the description of the coal seam position. By using this method to identify the coal seam in the coal seam profile image, the position and boundary of the coal seam can be accurately identified without a large amount of adjustment of the detection parameters, which reduces the complexity of the coal seam detection operation and improves the detection efficiency.

[0030] It is understandable that the fault image encoder and the coal seam image encoder can extract fault features and coal seam features using a neural network model trained with a coal seam-fault information data set consisting of a large number of coal seam profile images with differentially labeled coal seam and fault information. Figure 2 In this embodiment, the fault image encoder is an image encoder based on convolutional neural network and Focal-Transformer, and the coal seam image encoder is an image encoder based on ViTDet model. The local features of the fault are extracted from the coal seam profile image through the convolutional neural network algorithm and the Focal-Transformer algorithm, respectively, and the local features and global features of the coal seam are extracted from the coal seam profile image through the ViTDet model.

[0031] In this embodiment, the step of calculating the weights of the characteristics of each fault layer and the characteristics of the coal seam through the attention mechanism specifically includes the following steps: First, define a learnable L vector: L = {l1, l2, ..., l k}; Then the extracted fault features are used as the query vector Query, and the extracted coal seam features are used as the queried vector Key and content vector Value, that is, Q = F fault , K=F coal and V = F coal ; Finally, the weighted combination of fault and coal seam characteristics is calculated by the following formula:

[0032] F fused =α·V (1)

[0033]

[0034] Among them, SoftMax is the normalization function.

[0035] Reference Figure 2 In this embodiment, after weighting the features of coal seams and faults of each layer according to the calculated weighted combination of coal seams and faults, the weighted coal seam and fault features are fused and converted into intermediate representations through a Transformer-based feature decoder to obtain the final features for identifying the specific locations and boundaries of coal seams and faults, and the final features are passed to the next layer through the FPN network. The final features output by the last layer are passed to the R-CNN network through the FPN network for classification, and the final coal seam and fault detection results are output.

[0036] Based on the same inventive concept, an embodiment of the present invention further provides a computer-readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, the above-mentioned coal seam identification method can be implemented.

[0037] In some possible embodiments, various aspects of the coal seam identification method provided by the present invention can also be implemented in the form of a program product, which includes a program code. When the program product is run on an apparatus, the program code is used to enable the control device to execute the steps of the coal seam identification method according to various exemplary embodiments of the present application described above in this specification.

[0038] Based on the same inventive concept, an embodiment of the present invention also provides an identification device for implementing the above-mentioned coal seam identification method, including a processor and a memory, the memory is electrically connected to the processor, and the processor is used to execute the computer program stored in the memory to implement the above-mentioned coal seam identification method.

[0039] In one possible design, the processor may include one or more processing units, and the processor and memory may be implemented on the same chip or on separate chips. The processor may be a general-purpose processor, such as a central processing unit (CPU), a digital signal processor, an application-specific integrated circuit, a field programmable gate array or other programmable logic device, a discrete gate or transistor logic device, or a discrete hardware component, which may implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of the present application. The general-purpose processor may be a microprocessor or any conventional processor, etc. The steps of the coal seam identification method disclosed in the embodiments of the present application may be directly embodied as being executed by a hardware processor, or may be executed by a combination of hardware and software modules in the processor.

[0040] As a non-volatile computer-readable storage medium, the memory can be used to store non-volatile software programs, non-volatile computer executable programs and modules. The memory can include at least one type of storage medium, for example, it can include flash memory, hard disk, multimedia card, card-type memory, random access memory (Random Access Memory, RAM), static random access memory (Static RandomAccess Memory, SRAM), programmable read-only memory (Programmable Read Only Memory, PROM), read-only memory (Read Only Memory, ROM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, EEPROM), magnetic memory, disk, CD, etc. The memory is 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 in the embodiment of the present application can also be a circuit or any other device that can realize a storage function, for storing program instructions and / or data.

[0041] By designing and programming the processor, the code corresponding to the coal seam identification method described in the above embodiment can be fixed into the chip, so that the chip can execute the steps of the coal seam identification method of the embodiment shown in the present invention when running. How to design and program the processor is a technology well known to those skilled in the art and will not be described in detail here.

[0042] Based on the same inventive concept, an embodiment of the present invention also provides another coal seam identification device, including a fault image decoder, a coal seam image decoder, an attention mechanism module, a feature fusion module and a classification module. The fault image decoder is used to extract the local features of the fault in the coal seam profile image. The coal seam image decoder is used to extract the local features and global features of the coal seam in the coal seam profile image. The attention mechanism module is used to calculate the weights of the features of each layer of faults and coal seams through the attention mechanism, and weight the extracted features of the faults and coal seams. The feature fusion module is used to fuse the features of the coal seams and the features of the faults, and convert the fused features into the final features for identification. The classification module is used to classify the objects in the coal seam profile image according to the final features, and identify the specific positions and boundaries of the coal seams and faults.

[0043] In this embodiment, the fault image encoder is an image encoder based on a convolutional neural network and a Focal-Transformer, the coal seam image encoder is an image encoder based on a ViTDet model, and the feature fusion module is a feature decoder based on a Transformer. When calculating the weights of the features of the coal seam and the fault, the attention mechanism module uses the features of the fault as the Query of the attention mechanism, and uses the features of the coal seam as the Key and Value of the attention mechanism to calculate the weights of the features of the coal seam and the fault.

[0044] The present application is described with reference to the flowcharts and / or block diagrams of the methods, apparatus (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0045] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0046] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0047] In the description of the present invention, unless otherwise clearly defined, terms such as setting, installing, connecting, etc. should be understood in a broad sense, and technicians in the relevant technical field can reasonably determine the specific meanings of the above terms in the present invention based on the specific content of the technical solution.

[0048] The above-mentioned embodiments are only preferred embodiments of the present invention and cannot be used to limit the scope of protection of the present invention. Any non-substantial changes and substitutions made by technicians in this field on the basis of the present invention shall fall within the scope of protection required by the present invention.

Claims

1. A coal seam identification method, characterized in that: The steps include: Extracting local features of faults in coal seam profile images through the trained fault image encoder; Extracting local features and global features of the coal seam in the coal seam profile image through the trained coal seam image encoder; The weights of the features of each fault and coal seam are calculated through the attention mechanism, and the extracted features of the fault and coal seam are weighted; The coal seam features and the fault features are fused, and the fused features are converted into final features for identification; The objects in the coal seam profile image are classified according to the final features, and the specific locations and boundaries of the coal seams and faults are identified.

2. The coal seam identification method according to claim 1, characterized in that: In the step of extracting the local features of the fault in the coal seam profile image by using the trained fault image encoder, the local features of the fault are extracted by using an image encoder based on a convolutional neural network and a Focal-Transformer.

3. The coal seam identification method according to claim 1, characterized in that: In the step of extracting the local features and global features of the coal seam in the coal seam profile image by using the trained coal seam image encoder, the local features and global features of the coal seam are extracted by using an image encoder based on the ViTDet model.

4. The coal seam identification method according to claim 1, characterized in that: In the step of calculating the weights of the features of each layer of faults and coal seams through the attention mechanism and weighting the extracted features of the faults and coal seams, the features of the faults are used as the query of the attention mechanism, and the features of the coal seams are used as the key and value of the attention mechanism to calculate the weights of the features of each layer of faults and coal seams.

5. The coal seam identification method according to claim 1, characterized in that: In the step of fusing the coal seam features and the fault features and converting the fused features into final features for identification, the coal seam features and the fault features are fused through a Transformer-based feature decoder.

6. A storage medium storing a computer program, characterized in that: When the computer program is called and executed by the processor, the coal seam identification method according to any one of claims 1 to 5 is implemented.

7. A coal seam identification device, characterized in that: It comprises a processor and a memory, the processor is electrically connected to the memory, and the processor can implement the coal seam identification method according to any one of claims 1 to 5 by calling and executing a computer program in the memory.

8. A coal seam identification device, characterized in that: include: A fault image decoder, used to extract local features of faults in coal seam profile images; A coal seam image decoder, used to extract local and global features of the coal seam in the coal seam profile image; An attention mechanism module is used to calculate the weights of the features of each fault and coal seam through the attention mechanism, and to weight the features of the extracted faults and coal seams; A feature fusion module is used to fuse the features of the coal seam and the fault, and convert the fused features into final features for identification; The classification module is used to classify the objects in the coal seam profile image according to the final features and identify the specific locations and boundaries of the coal seams and faults.

9. The coal seam identification device according to claim 8, characterized in that: The tomographic image encoder is an image encoder based on a convolutional neural network and a Focal-Transformer, the coal seam image encoder is an image encoder based on a ViTDet model, and the feature fusion module is a feature decoder based on a Transformer.

10. The coal seam identification device according to claim 8, characterized in that: The attention mechanism module calculates the weights of the features of each fault and coal seam by using the features of the fault as the query of the attention mechanism and the features of the coal seam as the key and value of the attention mechanism.

Images