Lithology classification method and system, electronic equipment and storage medium

Abstract

The invention relates to the technical field of geological big data and mathematical geology, and particularly discloses a lithology classification method and system, electronic equipment and a storage medium, and the method comprises the steps: obtaining a target multispectral remote sensing image and a geochemical anomaly image of a target region, connecting the extracted first image feature of the target multispectral remote sensing image with the second image feature of the geochemical anomaly image to obtain a combined feature vector; and inputting the combined feature vector into a trained mixed attention module to obtain a target feature map, and inputting the target feature map into a KAN network to obtain a lithology classification result of each pixel point in the target feature map. According to the method, the multi-source geological data can be analyzed and processed, so that the characteristics of different types of geological data are extracted, automatic classification of the types of surface lithology is realized, and the accuracy of lithology classification is improved.

Description

Background Art

[0002] As a typical data-intensive discipline, geology includes remote sensing data, geochemical anomaly data, geophysical anomaly data, drilling data, and basic geological data, which have different structures and sources. These multi-source data have their own advantages in practical geological applications: remote sensing image data has the advantages of wide coverage, fast update and acquisition speed, and high spatial resolution; geochemical anomaly data has the advantage of higher sensitivity to shallow surface features and can more accurately reflect the mineral content information on the surface; geophysical anomaly data such as aeromagnetic and gravity have stronger penetration ability and can provide more effective reflection of deep hidden minerals. In summary, different types of geological data have their own characteristics. In the face of practical problems and applications, the discrimination results based on a single data source are difficult to achieve the application goals. Therefore, in the face of massive multi-source geological big data, it is necessary to unify the multi-source data format, integrate the advantages of multi-source data, and fuse multiple geological data to deal with the complex problems in practical applications.

[0003] At present, in the field of multi-source geological data fusion, common data processing methods include knowledge-driven based on prior knowledge or physical models and data-driven based on machine learning or deep learning. Among them, the knowledge-driven method mainly relies on the experience, theoretical knowledge and existing laws of domain experts, and usually solves problems based on prior knowledge and logical reasoning using known models and rules. The knowledge-driven method does not require a large amount of data to train the model, and the conclusions obtained are highly interpretable, but it consumes a lot of time and manpower, and the migration effect for new problems is also poor; on the other hand, the data-driven method mainly relies on a large amount of data, and discovers patterns and laws through data analysis and machine learning algorithms, so as to make predictions and decisions. Although data-driven can adapt to new problems as quickly as possible through model adjustment, and can efficiently extract rich information from the data itself, data-driven faces the "black box" problem of poor interpretability.

[0004] In summary, the current technical methods have the following problems: First, the classification effect based on a single data source is poor compared to the collaborative discrimination of multi-source data; second, the knowledge-driven and data-driven methods each have their own advantages and disadvantages, and cannot make up for each other's shortcomings and improve the accuracy of classification problems. Therefore, it is urgent to provide a technical solution to solve the above problems. Summary of the invention

[0005] In order to solve the above technical problems, the present invention provides a lithology classification method, system, electronic device and storage medium.

[0006] In a first aspect, the present invention provides a lithology classification method, the technical solution of which is as follows:

[0007] Acquire a target multispectral remote sensing image and a geochemical anomaly image of a target area, and connect the extracted first image feature of the target multispectral remote sensing image with the second image feature of the geochemical anomaly image to obtain a combined feature vector;

[0008] The combined feature vector is input into a trained hybrid attention module to obtain a target feature map, and the target feature map is input into a KAN network to obtain a lithology classification result for each pixel in the target feature map.

[0009] The beneficial effects of a lithology classification method of the present invention are as follows:

[0010] The method of the present invention can analyze and process multi-source geological data, thereby extracting features between geological data of different categories, realizing automatic classification of surface lithology types, and improving the accuracy of lithology classification.

[0011] Based on the above scheme, a lithology classification method of the present invention can also be improved as follows.

[0012] In an optional manner, the method further includes:

[0013] The original multispectral remote sensing image of the target area is acquired, and the original multispectral remote sensing image is subjected to image preprocessing to obtain the target multispectral remote sensing image; wherein the image preprocessing method includes at least one of radiation calibration, atmospheric correction, geometric correction and image cropping.

[0014] In an optional manner, the method further includes:

[0015] Acquire original geochemical sampling point data of the target area, and perform denoising and standardization processing on the original geochemical sampling point data to obtain target geochemical sampling point data;

[0016] A component balance analysis is performed on the target geochemical sampling point data to obtain a geochemical element combination, and Kriging interpolation is performed on the geochemical element combination to obtain the geochemical anomaly image.

[0017] In an optional manner, the method further includes:

[0018] The first image feature of the target multispectral remote sensing image and the second image feature of the geochemical anomaly image are acquired by using a feature extraction network.

[0019] In an optional manner, the hybrid attention module includes: a channel attention module and a spatial attention module.

[0020] In a second aspect, the present invention provides a lithology classification system, the technical solution of which is as follows:

[0021] Includes: processing unit and operation unit;

[0022] The processing unit is used to: obtain a target multispectral remote sensing image and a geochemical anomaly image of a target area, and connect the extracted first image feature of the target multispectral remote sensing image with the second image feature of the geochemical anomaly image to obtain a combined feature vector;

[0023] The operation unit is used to: input the combined feature vector into a trained hybrid attention module to obtain a target feature map, and input the target feature map into a KAN network to obtain a lithology classification result for each pixel in the target feature map.

[0024] The beneficial effects of a lithology classification system of the present invention are as follows:

[0025] The system of the present invention can analyze and process multi-source geological data, thereby extracting features between different categories of geological data, realizing automatic classification of surface lithology types, and improving the accuracy of lithology classification.

[0026] Based on the above scheme, a lithology classification system of the present invention can also be improved as follows.

[0027] In an optional manner, the method further includes: a first preprocessing unit; the first preprocessing unit is used to:

[0028] The original multispectral remote sensing image of the target area is acquired, and the original multispectral remote sensing image is subjected to image preprocessing to obtain the target multispectral remote sensing image; wherein the image preprocessing method includes at least one of radiation calibration, atmospheric correction, geometric correction and image cropping.

[0029] In an optional manner, the method further includes: a second preprocessing unit; the second preprocessing unit is used to:

[0030] Acquire original geochemical sampling point data of the target area, and perform denoising and standardization processing on the original geochemical sampling point data to obtain target geochemical sampling point data;

[0031] A component balance analysis is performed on the target geochemical sampling point data to obtain a geochemical element combination, and Kriging interpolation is performed on the geochemical element combination to obtain the geochemical anomaly image.

[0032] In a third aspect, a technical solution of an electronic device of the present invention is as follows:

[0033] The invention comprises a memory, a processor and a program stored in the memory and running on the processor. When the processor executes the program, the steps of the lithology classification method of the present invention are realized.

[0034] In a fourth aspect, the present invention provides a computer-readable storage medium having the following technical solution:

[0035] Instructions are stored in the computer-readable storage medium. When the computer-readable storage medium reads the instructions, the computer-readable storage medium executes the steps of the lithology classification method of the present invention.

[0036] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented according to the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] The accompanying drawings are only used to illustrate the embodiments and are not to be considered as limiting the present invention. In addition, the same reference symbols are used to represent the same components throughout the accompanying drawings. In the accompanying drawings:

[0038] Figure 1 A schematic diagram of a process flow of an embodiment of a lithology classification method of the present invention;

[0039] Figure 2 is a schematic diagram of a target multispectral remote sensing image;

[0040] Figure 3 It is a schematic diagram of geochemical anomaly images;

[0041] Figure 4 It is a structural diagram of the feature extraction network;

[0042] Figure 5 Schematic diagram of the structure of the channel attention module;

[0043] Figure 6 Schematic diagram of the structure of the spatial attention module;

[0044] Figure 7 A structural schematic diagram of an embodiment of a lithology classification system of the present invention;

[0045] Figure 8 The figure is a schematic structural diagram of an embodiment of an electronic device of the present invention. DETAILED DESCRIPTION

[0046] Exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be limited to the embodiments set forth herein.

[0047] Figure 1 A flow chart of an embodiment of a lithology classification method provided by the present invention is shown, and the lithology classification method can be executed by an electronic device such as a terminal device or a server. The terminal device can be any fixed or mobile terminal such as a user equipment (UE), a mobile device, a user terminal, a terminal, a cellular phone, a cordless phone, a personal digital assistant (PDA), a handheld device, a computing device, a vehicle-mounted device, a wearable device, etc. The server can be a single server or a server cluster composed of multiple servers. Any electronic device can implement the lithology classification method by calling computer-readable instructions stored in a memory through a processor. Figure 1 As shown, the following steps are included:

[0048] S1. Acquire a target multispectral remote sensing image and a geochemical anomaly image of a target area, and connect the extracted first image feature of the target multispectral remote sensing image with the second image feature of the geochemical anomaly image to obtain a combined feature vector.

[0049] Among them, the target area is an arbitrarily selected area, which can be selected according to actual needs and is not limited here. The first image feature is the image feature of the target multispectral remote sensing image, and the second image feature is the image feature of the geochemical anomaly image. Using the feature extraction network, the first image feature of the target multispectral remote sensing image and the second image feature of the geochemical anomaly image are obtained respectively. Through the concat layer, the first image feature and the second image feature are connected to obtain a combined feature vector.

[0050] 1) Before S1, it also includes:

[0051] An original multispectral remote sensing image of the target area is acquired, and image preprocessing is performed on the original multispectral remote sensing image to obtain the target multispectral remote sensing image.

[0052] Among them, the image preprocessing methods include: at least one of radiation calibration, atmospheric correction, geometric correction and image cropping. The original multispectral remote sensing image is the ASTER multispectral remote sensing image directly collected. ASTER contains 3 visible near-infrared bands (0.52-0.86μm) with a spatial resolution of 15m, 6 short-wave infrared bands (1.60-2.43μm) with a spatial resolution of 30m, and 5 thermal infrared bands (8.13-11.65μm) with a spatial resolution of 90m. Each ASTER scene obtains a 60km strip and is subdivided into 60×60km. 2 scene. Among them, bands 1, 2, and 3 are selected with a spatial resolution of 15m for subsequent processing. The image preprocessing in this embodiment includes: radiation calibration, atmospheric correction, geometric correction, and image cropping. Specifically: according to the size of the target area, the original multispectral remote sensing image is cropped to obtain a first image; the first image is geometrically corrected to obtain a second image after geometric correction; the second image is radiometrically calibrated to obtain a third image after radiation calibration; the third image is atmospherically corrected to obtain a fourth image after atmospheric correction; the fourth image is determined as the target multispectral remote sensing image (such as Figure 2 shown).

[0053] 2) Before S1, it also includes:

[0054] The original geochemical sampling point data of the target area is obtained, and the original geochemical sampling point data is denoised and standardized to obtain the target geochemical sampling point data.

[0055] Among them, the original geochemical sampling point data is the geochemical sampling point data collected directly. Geochemical sampling point data can reveal the geological composition and characteristics of different regions of the earth, including the element content, distribution and composition of rocks, soils, water bodies, etc., which helps to understand the internal structure and evolution of the earth. On the other hand, geochemical sampling point data can be used to monitor pollutants in the environment, such as heavy metals, organic matter, etc., and to assess environmental quality and ecosystem health. According to the 1:20w geochemical anomaly map standard, soil sample content data of 39 geochemical elements or compounds such as Ag, As, Au, B, Ba, Cu, Mn, Mo, Al2O3, CaO, Fe2O3, and MgO in the target area are uniformly sampled at equal intervals.

[0056] The denoising process is as follows: screening the original geochemical sampling point data for abnormal points and removing abnormal outlier points; the standardization process is as follows: standardizing the element data in the original geochemical sampling point data.

[0057] Perform component balance analysis on the target geochemical sampling point data to obtain a geochemical element combination, and perform Kriging interpolation on the geochemical element combination to obtain the geochemical anomaly image. Specifically:

[0058] ① Conduct compositional balance analysis on the target geochemical sampling point data to obtain its strongly correlated geochemical element combination (multiple in number).

[0059] ② Analyze relevant geochemical combinations of target geochemical sampling data based on prior knowledge, and preliminarily divide all geochemical elements into groups.

[0060] ③ The SBP binary partitioning method is used for the geochemical data after the above grouping to divide the groups in the data set into non-overlapping subgroups until all elements are divided into different partitions, and the equilibrium value b is calculated for each partition (group). i : and represents the number of "+" and "-" in the i-th partition, and It represents the geometric mean of its corresponding part. "+" and "-" refer to the numerator and denominator of the partition respectively.

[0061] ④ Determine the partition (group) with the largest balance value as the geochemical element combination (the quantity is one).

[0062] ⑤ Perform Kriging interpolation on the geochemical element combination to obtain geochemical anomaly images (such as Figure 3 shown).

[0063] 3) Using a feature extraction network, obtaining the first image feature of the target multispectral remote sensing image and the second image feature of the geochemical anomaly image.

[0064] Among them, Figure 4As shown, the feature extraction network includes: a convolution layer, a maximum pooling layer, a first basic block module and a second basic block module connected in sequence. The first basic block module includes: a 5×5 convolution layer, a batch normalization layer, a ReLU as an activation function and a dropout layer connected in sequence. The dropout layer is used as a regularization term to improve the generalization ability of the model and to prevent overfitting. The size of the input data (target multispectral remote sensing image or geochemical anomaly image) is 64×64×3, the batch size is 32, and the output feature size after passing through the above feature extraction network is 8×8×128. Therefore, the size of the output feature set (combined feature vector) F after the two data sources (first image features and second image features) are finally connected in the channel dimension is 8×8×256.

[0065] S2. Input the combined feature vector into a trained hybrid attention module to obtain a target feature map, and input the target feature map into a KAN network to obtain a lithology classification result for each pixel in the target feature map.

[0066] The hybrid attention module includes: channel attention module and spatial attention module. The KAN network replaces the MLP multi-layer perceptron network to achieve lithology classification of each pixel.

[0067] like Figure 5 As shown in the figure, the processing process of the channel attention module includes: ① The combined feature vector F(H×W×C) is passed through a maximum pooling layer and an average pooling layer respectively to obtain two 1×1×C feature maps: Maxpool(F) and Avepool(F). ② The two feature maps (Maxpool(F) and Avepool(F)) are respectively input into a shared weight neural network layer (MLP) structure. The MLP includes two fully connected layers and one hidden layer to calculate the weight corresponding to each channel. ③ The feature maps MLP(Maxpool(F)) and MLP(Avepool(F)) processed by MLP are added element by element, and activated by the Sigmoid function to obtain the weight correction matrix M in the channel dimension. C (F), that is: M C (F) = σ{MLP[Maxpool(F)]+MLP[Avepool(F)]}; ④ Correct the weight matrix M on the channel dimension C (F) is multiplied element by element with the combined feature vector F to obtain the feature map F after channel attention weight allocation C .

[0068] like Figure 6As shown in the figure, the processing process of the spatial attention module includes: ① The feature map F output by the channel attention module C Input the maximum pooling layer and the average pooling layer respectively, and perform pooling in the channel dimension to obtain two H×W×1 feature maps (Maxpool(F C ) and Avepool(F C )) and splice them in the channel dimension. ② The spliced ​​feature map is weighted by 3×3 convolution and output through the Sigmoid function to obtain the weight correction matrix M in the spatial dimension S (F C ), that is: M S (F C )=σ{Conv 3×3 [Maxpool(F C ); Avepool(F C )]}; ③ Modify the weight matrix M in the spatial dimension S (F C ) and the feature map F output by the channel attention module C F C By performing element-by-element multiplication, we can obtain the target feature map F′ corrected by the spatial attention mechanism.

[0069] It should be noted that dataset labels are made, hybrid attention models are trained, and surface lithology is classified. The training dataset is channel normalized according to different lithologies, and then the dataset is cropped into patches according to the sliding window method, with a size of 64*64 and a sliding step of 32. Finally, 4 lithologies and 4 channels of tif images are obtained, totaling 4*4*5089=81424 tif data. They are shuffled and divided into the dataset according to the ratio of training data: verification data: test data = 70%: 15%: 15%. The hybrid attention model constructed above is trained based on the training data according to the back propagation algorithm, and the cross entropy loss function is used, which is defined as follows: y i represents the real image component, It represents the decomposed image components, minimizes the difference between the two, and uses the NesterovAdam (Nadam) optimizer to make the loss function tend to the global minimum. The initial learning rate is set to 0.0001, the decay rate is (0.99, 0.99), and the weight matrix of the output feature vector is obtained.

[0070] The lithology classification method based on hybrid drive proposed in the technical solution of this embodiment overcomes the differences between knowledge-driven and data-driven methods and integrates the advantages of the two research methods. At the same time, the advantages of two different data are combined, the sensitivity of geochemical data to element content and the high spatial resolution advantage of remote sensing data are integrated, and the accuracy of lithology classification is improved. This embodiment eliminates the closure effect between geochemical elements through component balance analysis, reduces the error caused by the data closure effect, and at the same time, the component balance analysis selects the geochemical element combination related to lithology classification among many geochemical elements as knowledge-driven prior knowledge. This embodiment further allocates weights to local features by introducing feature extraction network Resnet and hybrid attention mechanism, and captures the most important information of the two images. This method reduces the dimension of features, increases the information density of data, retains the accuracy of geochemical data, and also improves the spatial resolution of geochemical data in combination with remote sensing data, and achieves good results in lithology classification.

[0071] The technical solution of this embodiment can analyze and process multi-source geological data, thereby extracting features between different categories of geological data, realizing automatic classification of surface lithology types, and improving the accuracy of lithology classification.

[0072] Figure 7 FIG. 2 is a schematic diagram showing a structural example of an embodiment of a lithology classification system 200 provided by the present invention. Figure 7 As shown, the system 200 includes: a processing unit 210 and an operation unit 220;

[0073] The processing unit 210 is used to: obtain a target multispectral remote sensing image and a geochemical anomaly image of a target area, and connect the extracted first image feature of the target multispectral remote sensing image with the second image feature of the geochemical anomaly image to obtain a combined feature vector;

[0074] The operation unit 220 is used to: input the combined feature vector into a trained hybrid attention module to obtain a target feature map, and input the target feature map into a KAN network to obtain a lithology classification result for each pixel in the target feature map.

[0075] In an optional manner, the method further includes: a first preprocessing unit; the first preprocessing unit is used to:

[0076] The original multispectral remote sensing image of the target area is acquired, and the original multispectral remote sensing image is subjected to image preprocessing to obtain the target multispectral remote sensing image; wherein the image preprocessing method includes at least one of radiation calibration, atmospheric correction, geometric correction and image cropping.

[0077] In an optional manner, the method further includes: a second preprocessing unit; the second preprocessing unit is used to:

[0078] Acquire original geochemical sampling point data of the target area, and perform denoising and standardization processing on the original geochemical sampling point data to obtain target geochemical sampling point data;

[0079] A component balance analysis is performed on the target geochemical sampling point data to obtain a geochemical element combination, and Kriging interpolation is performed on the geochemical element combination to obtain the geochemical anomaly image.

[0080] In an optional manner, the method further includes: an extraction unit; the extraction unit is used to:

[0081] The first image feature of the target multispectral remote sensing image and the second image feature of the geochemical anomaly image are acquired by using a feature extraction network.

[0082] In an optional manner, the hybrid attention module includes: a channel attention module and a spatial attention module.

[0083] The technical solution of this embodiment can analyze and process multi-source geological data, thereby extracting features between different categories of geological data, realizing automatic classification of surface lithology types, and improving the accuracy of lithology classification.

[0084] The above parameters and steps for each module to implement corresponding functions in the lithology classification system 200 of this embodiment can refer to the parameters and steps in the embodiment of the lithology classification method above, and will not be repeated here.

[0085] like Figure 8 As shown, an electronic device 300 according to an embodiment of the present invention includes a processor 320, the processor 320 is coupled to a memory 310, and the memory 310 stores at least one computer program 330, and the at least one computer program 330 is loaded and executed by the processor 320, so that the electronic device 300 implements any of the above-mentioned lithology classification methods, specifically:

[0086] The electronic device 300 may have relatively large differences due to different configurations or performances, and may include one or more processors 320 (Central Processing Units, CPU) and one or more memories 310, wherein at least one computer program 330 is stored in the one or more memories 310, and the at least one computer program 330 is loaded and executed by the one or more processors 320, so that the electronic device 300 implements any lithology classification method provided in the above embodiments. Of course, the electronic device 300 may also have components such as a wired or wireless network interface, a keyboard, and an input and output interface for input and output, and the electronic device 300 may also include other components for realizing device functions, which will not be described in detail here.

[0087] A computer-readable storage medium according to an embodiment of the present invention stores at least one computer program, and the at least one computer program is loaded and executed by a processor so that a computer implements any one of the above-mentioned lithology classification methods.

[0088] Optionally, the computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a compact disc (CD-ROM), a magnetic tape, a floppy disk, an optical data storage device, and the like.

[0089] In an exemplary embodiment, a computer program product or a computer program is also provided, the computer program product or the computer program comprising computer instructions, the computer instructions being stored in a computer-readable storage medium. A processor of an electronic device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the electronic device performs any of the above-mentioned lithology classification methods.

[0090] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application are used to distinguish similar objects and represent the definition of a specific order or sequence. The order of use of similar objects can be interchanged where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than the order shown or described.

[0091] Those skilled in the art know that the present invention can be implemented as a system, method or computer program product. Therefore, the present disclosure can be specifically implemented in the following forms, namely: it can be complete hardware, it can be complete software (including firmware, resident software, microcode, etc.), or it can be a combination of hardware and software, which is generally referred to as "circuit", "module" or "system" herein. In addition, in some embodiments, the present invention can also be implemented in the form of a computer program product in one or more computer-readable media, and the computer-readable medium contains computer-readable program code.

[0092] Any combination of one or more computer-readable media may be used. A computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in conjunction with an instruction execution system, device, or device.

[0093] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention. A person skilled in the art may change, modify, replace and vary the above embodiments within the scope of the present invention.

Claims

1. A lithology classification method, characterized in that: include: Acquire a target multispectral remote sensing image and a geochemical anomaly image of a target area, and connect the extracted first image feature of the target multispectral remote sensing image with the second image feature of the geochemical anomaly image to obtain a combined feature vector; The combined feature vector is input into a trained hybrid attention module to obtain a target feature map, and the target feature map is input into a KAN network to obtain a lithology classification result for each pixel in the target feature map.

2. The lithology classification method according to claim 1, characterized in that: Also includes: The original multispectral remote sensing image of the target area is acquired, and the original multispectral remote sensing image is subjected to image preprocessing to obtain the target multispectral remote sensing image; wherein the image preprocessing method includes at least one of radiation calibration, atmospheric correction, geometric correction and image cropping.

3. The lithology classification method according to claim 1, characterized in that: Also includes: Acquire original geochemical sampling point data of the target area, and perform denoising and standardization processing on the original geochemical sampling point data to obtain target geochemical sampling point data; A component balance analysis is performed on the target geochemical sampling point data to obtain a geochemical element combination, and Kriging interpolation is performed on the geochemical element combination to obtain the geochemical anomaly image.

4. The lithology classification method according to claim 1, characterized in that: Also includes: The first image feature of the target multispectral remote sensing image and the second image feature of the geochemical anomaly image are acquired by using a feature extraction network.

5. The lithology classification method according to any one of claims 1 to 4, characterized in that: The hybrid attention module includes: a channel attention module and a spatial attention module.

6. A lithology classification system, characterized in that: include: Processing units and operation units; The processing unit is used to: obtain a target multispectral remote sensing image and a geochemical anomaly image of a target area, and connect the extracted first image feature of the target multispectral remote sensing image with the second image feature of the geochemical anomaly image to obtain a combined feature vector; The operation unit is used to: input the combined feature vector into a trained hybrid attention module to obtain a target feature map, and input the target feature map into a KAN network to obtain a lithology classification result for each pixel in the target feature map.

7. The lithology classification system according to claim 6, characterized in that: Also includes: a first pre-processing unit; The first pre-processing unit is used for: The original multispectral remote sensing image of the target area is acquired, and the original multispectral remote sensing image is subjected to image preprocessing to obtain the target multispectral remote sensing image; wherein the image preprocessing method includes at least one of radiation calibration, atmospheric correction, geometric correction and image cropping.

8. The lithology classification system according to claim 6, characterized in that: Also includes: A second pre-processing unit; The second pre-processing unit is used for: Acquire original geochemical sampling point data of the target area, and perform denoising and standardization processing on the original geochemical sampling point data to obtain target geochemical sampling point data; A component balance analysis is performed on the target geochemical sampling point data to obtain a geochemical element combination, and Kriging interpolation is performed on the geochemical element combination to obtain the geochemical anomaly image.

9. An electronic device, characterized in that: The electronic device includes a processor, the processor is coupled to a memory, the memory stores at least one computer program, and the at least one computer program is loaded and executed by the processor so that the electronic device implements the lithology classification method as described in any one of claims 1 to 5.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores at least one computer program, and the at least one computer program is loaded and executed by the processor so that the computer-readable storage medium implements the lithology classification method according to any one of claims 1 to 5.

Images