Remote sensing image semantic segmentation method and device based on wavelet transform convolution

Abstract

The invention discloses a remote sensing image semantic segmentation method and device based on wavelet transform convolution. The method comprises the following steps: firstly, extracting semantic feature information in a remote sensing image based on a wavelet transform convolution encoder; secondly, the encoder and the decoder are connected through double-cross attention, the attention uses channel cross attention and space cross attention through series connection, feature information output by the encoder is enhanced, the validity of jump connection between the encoder and the decoder is enhanced, and the semantic gap between the encoder and the decoder is reduced; and finally, a traditional CNN decoder is used for the feature map so as to obtain a prediction result, and the prediction result is the semantic segmentation node of the remote sensing image. According to the method, semantic context information in the remote sensing image can be efficiently captured, the response of the model to shape features is enhanced, and the semantic segmentation precision is improved.

Description

Technical Field

[0001] The invention belongs to the technical field of remote sensing image processing, and in particular relates to a remote sensing image semantic segmentation method and device based on wavelet transform convolution. Background Art

[0002] Semantic segmentation of remote sensing images is a key task in the field of computer vision and remote sensing. It involves assigning each pixel in a remote sensing image to a predefined category label, such as city, forest, water body, etc. This technique is essential for a variety of applications such as environmental monitoring, urban planning, agricultural management, and disaster assessment. Traditional methods for semantic segmentation of remote sensing images usually rely on manual feature extraction and classical machine learning algorithms. Traditional methods usually require expert knowledge and manual feature extraction, which is time-consuming and complex, and difficult to adapt to new datasets or changing environmental conditions. Moreover, they often have poor generalization ability and are difficult to adapt to the rich features of objects in modern high-precision remote sensing images.

[0003] Currently, the more mainstream methods are those based on deep learning. Deep learning, especially convolutional neural networks (CNNs), has become the core technology for semantic segmentation of remote sensing images. These models can automatically learn complex feature representations from large amounts of data without relying on manual feature extraction, which is an essential step in traditional methods. Deep learning models learn mappings directly from pixel-level labels through end-to-end training, improving segmentation accuracy and efficiency. Summary of the invention

[0004] In view of the efficiency problem of remote sensing semantic segmentation, the purpose of the present invention is to provide a remote sensing image semantic segmentation method and device based on wavelet transform convolution. By adopting wavelet transform convolution, the semantic context information in the features is captured, the distinguishability of ground object classification is enhanced, and the purpose of accurately extracting remote sensing image semantic segmentation targets is achieved.

[0005] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0006] In a first aspect, the present invention provides a remote sensing image semantic segmentation method based on wavelet transform convolution, comprising:

[0007] Acquire remote sensing images to be predicted;

[0008] Inputting the remote sensing image to be predicted into a pre-built remote sensing image semantic segmentation prediction model to obtain a semantic segmentation result of the remote sensing image; the semantic segmentation result of the remote sensing image refers to the classification of target objects for each pixel in the remote sensing image;

[0009] The remote sensing image semantic segmentation prediction model is constructed as follows:

[0010] Constructing a remote sensing image semantic segmentation network model; the remote sensing image semantic segmentation network model comprises: an encoder, a skip connection and a decoder; the encoder is used to extract a semantic information feature map of an input image, the skip connection is used to enhance the encoder output feature, and the decoder is used to output a prediction result; the prediction result refers to a prediction map with the same size as the input image, and the prediction map classifies the target object for each pixel in the remote sensing image;

[0011] A remote sensing image benchmark data set is obtained for label annotation and divided into a training set and a test set. The remote sensing image semantic segmentation network model is trained based on the training set to obtain a network model with optimal parameters as a remote sensing image semantic segmentation prediction model.

[0012] Preferably, before each iterative training, the images in the training set and the test set are subjected to random cropping, random horizontal sliding, random flipping, rotation and color enhancement operations.

[0013] Preferably, the remote sensing image semantic segmentation network model includes: four encoders, four skip connections and four decoders;

[0014] The encoder extracts a semantic information feature map of the input image based on wavelet transform convolution;

[0015] The skip connection is based on channel cross attention and spatial cross attention to enhance the semantic information feature map.

[0016] Preferably, the encoder extracts a semantic information feature map of the input image based on wavelet transform convolution, comprising:

[0017] Input remote sensing image use Convolution extracts features and obtains feature maps ;

[0018] For feature maps Using wavelet decomposition, let the initial low-pass component , intermediate features ,in is the initial weight, and four fluxes of different frequencies are obtained by cascade wavelet decomposition: , , which correspond to the semantic information feature maps obtained by encoders at different layers, and convolution operations are performed on fluxes of different frequencies to obtain four intermediate features of different frequencies: ;in, Refers to wavelet transform, is the convolution operation, is the weight.

[0019] Preferably, the skip connection is based on channel cross attention and spatial cross attention to enhance the semantic information feature map, including:

[0020] Intermediate features of the encoder output After block processing, we get , :

[0021] ,

[0022] in, represents a two-dimensional average pooling operation, Reinvention, Represents a one-dimensional depth-wise separable convolution;

[0023] right Perform channel cross attention processing and first generate Three variables, denoted as query, key, and value, are represented as:

[0024] , , ,

[0025] Finally, the feature map is obtained: ,

[0026] in, is a constant, represents channel cross attention, refer to Activation function;

[0027] For feature maps Perform spatial cross attention processing and first generate Three variables, denoted as query, key, and value, are represented as:

[0028] , , ;

[0029] Finally, the enhanced semantic information feature map is obtained: ,in, is a constant.

[0030] Preferably, the decoder outputs a prediction result, including:

[0031] Set mixed features ,for , ;

[0032] make ,get ; Where IWT stands for inverse wavelet transform,

[0033] In the first decoder, the obtained Up-sampling, Restore to original image size and use The convolution outputs the final prediction result.

[0034] Preferably, during the training process of the remote sensing image semantic segmentation network model, a cross entropy loss function is used to calculate the loss, and the model parameters are updated by back propagation.

[0035] In a second aspect, the present invention further provides a remote sensing image semantic segmentation device based on wavelet transform convolution, which is used to implement the above-mentioned remote sensing image semantic segmentation method based on wavelet transform convolution, and the device comprises:

[0036] A data acquisition module, used for acquiring remote sensing images to be predicted;

[0037] Remote sensing image semantic segmentation prediction model training module, used to train remote sensing image semantic segmentation prediction model;

[0038] A prediction output module, used for inputting the remote sensing image to be predicted into the remote sensing image semantic segmentation prediction model to obtain a semantic segmentation result of the remote sensing image; the remote sensing semantic segmentation result refers to the identification and classification of ground objects for each pixel in the remote sensing image;

[0039] The training process of the remote sensing image semantic segmentation prediction model training module is as follows:

[0040] Constructing a remote sensing image semantic segmentation network model; the remote sensing image semantic segmentation network model comprises: an encoder, a skip connection and a decoder; the encoder is used to extract a semantic information feature map of an input image, the skip connection is used to enhance the encoder output feature, and the decoder is used to output a prediction result; the prediction result refers to a prediction map with the same size as the input image, and the prediction map classifies the target object for each pixel in the remote sensing image;

[0041] A remote sensing image benchmark data set is obtained for label annotation and divided into a training set and a test set. The remote sensing image semantic segmentation network model is trained based on the training set to obtain a network model with optimal parameters as a remote sensing image semantic segmentation prediction model.

[0042] In a third aspect, the present invention provides a computer-readable storage medium storing one or more programs, wherein the one or more programs include instructions, which, when executed by a computing device, enable the computing device to perform any of the above-mentioned remote sensing image semantic segmentation methods based on wavelet transform convolution.

[0043] In a fourth aspect, the present invention provides a computing device comprising one or more processors, a memory and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include instructions for executing any of the above-mentioned remote sensing image semantic segmentation methods based on wavelet transform convolution.

[0044] The beneficial effects of the present invention are:

[0045] The present invention provides a remote sensing image semantic segmentation method based on wavelet transform, which constructs a remote sensing image semantic segmentation prediction model through an encoder-decoder structure. The remote sensing image semantic segmentation prediction model is formed by combining a traditional convolutional neural network (CNN) and wavelet transform. It can obtain a receptive field close to the global without significantly increasing parameters, thereby efficiently extracting semantic information in remote sensing images; secondly, the present invention uses double cross attention to connect the encoder and the decoder, which enhances the feature information output by the encoder and the effectiveness of the jump connection between the encoder and the decoder by using channel cross attention and spatial cross attention in series, and reduces the semantic gap between them; finally, the traditional CNN decoder is used for the feature map to obtain the prediction result. The present invention can overcome the difficulties brought by the difference in ground object classification targets, efficiently capture the semantic context information in remote sensing images, realize the accurate segmentation of ground object classification targets by the model, and improve the accuracy of semantic segmentation. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 Schematic diagram of the training process of the remote sensing image semantic segmentation prediction model in an embodiment of the present invention;

[0047] Figure 2 Schematic diagram of the remote sensing image semantic segmentation network model structure in an embodiment of the present invention;

[0048] Figure 3 It is a flow chart of a remote sensing image semantic segmentation method based on wavelet transform convolution provided by an embodiment of the present invention;

[0049] Figure 4 Schematic diagram of the data processing process of the wavelet transform convolution encoder in an embodiment of the present invention;

[0050] Figure 5 It is a schematic diagram of the double cross attention jump connection data processing process in an embodiment of the present invention. DETAILED DESCRIPTION

[0051] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments and the accompanying drawings. Here, the illustrative embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.

[0052] It should also be noted that, in order to avoid obscuring the present invention due to unnecessary details, only structures and / or processing steps closely related to the solutions according to the present invention are shown in the accompanying drawings, while other details that are not closely related to the present invention are omitted.

[0053] It should be emphasized that the term “include / comprises” when used herein refers to the presence of features, elements, steps or components, but does not exclude the presence or addition of one or more other features, elements, steps or components.

[0054] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. In the accompanying drawings, the same reference numerals represent the same or similar components, or the same or similar steps.

[0055] It should be emphasized here that the step marks mentioned below are not intended to limit the order of the steps, but it should be understood that the steps can be executed in the order mentioned in the embodiment, or in a different order from the embodiment, or several steps can be executed simultaneously.

[0056] The first embodiment of the present invention provides a remote sensing image semantic segmentation method based on wavelet transform convolution, comprising:

[0057] Acquire remote sensing images to be predicted;

[0058] The acquired remote sensing image is input into the pre-built remote sensing image semantic segmentation prediction model to obtain the semantic segmentation result of the remote sensing image.

[0059] It should be noted that the remote sensing image to be predicted is obtained in the following way:

[0060] This case is to perform semantic segmentation on high-resolution remote sensing images, such as those in the ISPRS Potsdam and Vaihingen datasets. Such remote sensing images can be obtained on UAV remote sensing platforms and contain target objects such as cars, trees, bushes, buildings, and hard roads, as well as other background objects.

[0061] It should be noted that the semantic segmentation results of remote sensing images include:

[0062] Remote sensing image semantic segmentation performs pixel-level image classification on remote sensing images. The result of remote sensing image semantic segmentation is a prediction map of the same size as the remote sensing image, which classifies the target object for each pixel in the image. The prediction map labels the target object pixels, and different colors represent different types of objects. For example, yellow represents cars, green represents trees, blue represents buildings, and red pixels represent backgrounds.

[0063] See also Figure 1In one embodiment, a remote sensing image semantic segmentation prediction model is pre-built, and the specific implementation method is as follows:

[0064] S1. Constructing a remote sensing image semantic segmentation network model; see Figure 2 ,The remote sensing image semantic segmentation network model includes: four encoders, four skip connections and four decoders;

[0065] Among them, the encoder is used to extract the semantic information feature map. In this embodiment, the encoder uses wavelet transform convolution, which expands the receptive field by combining wavelet transform, so that the convolutional neural network can obtain a receptive field close to the global without significantly increasing the parameters; the input is divided into different frequency bands by wavelet decomposition, allowing the convolution layer to process the low-frequency and high-frequency components separately, thereby enhancing the model's response to low-frequency components (i.e., shape features).

[0066] The jump connection is used to enhance the encoder output features. In this embodiment, the jump connection uses double cross attention, with the help of channel cross attention (CCA) and spatial cross attention (SCA), to enhance the effectiveness of the U-Net jump connection, capture richer global context information, and enable the model to better extract fine-grained features.

[0067] The decoder is used to process the encoder output features to obtain the prediction result, which is the semantic segmentation result of the remote sensing image. (The prediction result here refers to a prediction map with the same size as the input image, which classifies the target objects for each pixel in the remote sensing image. The prediction map generated by the prediction module is used to compare with the real label map to calculate the loss optimization model during the model training stage; the semantic segmentation result of the remote sensing image is generated during the model prediction stage).

[0068] S2. Obtain a remote sensing image benchmark dataset for labeling and divide it into a training set and a test set. Input the training set into the constructed remote sensing image semantic segmentation network model, extract and enhance the semantic feature information of the input image, and output the prediction result. The label refers to the classification of each pixel in the remote sensing image. The purpose of using the real label is to compare and calculate the optimization model with the model prediction result during the model training stage. The prediction result refers to a prediction map with the same size as the input image, which classifies the target object for each pixel in the remote sensing image.

[0069] S3. Based on the loss between the prediction result and the true label, the parameters of the constructed remote sensing image semantic segmentation network model are updated, and the network model with the optimal parameters is obtained through continuous iterative training until the maximum number of iterations is reached or the model reaches the convergence condition, which is used as the remote sensing image semantic segmentation prediction model;

[0070] S4. Use the test set to test the performance of the constructed remote sensing image semantic segmentation prediction model.

[0071] In this embodiment, the ISPRS Potsdam dataset is selected as the remote sensing image benchmark dataset.

[0072] It should be noted that before each iteration, the images in the training set and the test set are randomly cropped, randomly horizontally slid, randomly flipped, rotated, and color enhanced to enhance the generalization performance of the model.

[0073] It should be noted that during the training process, the cross entropy loss function is used to calculate the difference between the predicted result and the true label, and the model parameters are updated by back propagation to make the loss value as small as possible. Set the maximum number of iterations. When the training reaches the set maximum number of iterations, stop the training, save the best model parameters and import them into the network model to obtain the remote sensing image semantic segmentation prediction model. The cross entropy loss function is shown in the formula:

[0074] ,

[0075] in is the true label of the sample, which takes a value of 0 or 1. is the predicted probability. The cross entropy loss function measures the difference between the probability distribution predicted by the model and the probability distribution of the true label.

[0076] In one embodiment, the maximum number of iterations is set to 200.

[0077] In one embodiment, the asymmetric semantic enhancement module is based on asymmetric convolution and global average pooling operations. The global average pooling operation can integrate global information into the semantic context information to improve segmentation accuracy; while the asymmetric convolution can reduce the amount of calculation without changing the size of the receptive field, thereby improving calculation efficiency.

[0078] See also Figure 4 , the process of extracting semantic information feature map by wavelet transform convolution encoder is as follows ( The value of is 1 to 4, because the model of the present invention has 4 layers):

[0079] 11) Assume that the input remote sensing image is .in, are the height, width and number of channels of the feature map respectively. use Convolution extracts features and obtains feature maps ;

[0080] 12) Feature map Using wavelet decomposition, let the initial low-pass component , intermediate features ,in is the initial weight, which is a randomly generated set of 0~1 values. The intermediate feature refers to the feature that will be passed to the decoder through the jump connection processing; , four fluxes of different frequencies are obtained by cascading wavelet decomposition, (corresponding to the semantic information feature maps obtained at different layers in the encoder), and convolution operations are performed on fluxes of different frequencies to obtain four intermediate features of different frequencies ;

[0081] in It refers to wavelet transform, and the specific operation is to first give a low-pass filter , a set of high-pass filters , and then use these four filters as weights for the input Perform convolution operation ( ), and obtain four components (among which, is the low-pass component, Represents the high frequency components of horizontal, vertical and diagonal lines respectively) .

[0082] The advantage of wavelet transform is that it can extract the low-pass components (i.e., shape features) in the image, and then the low-pass components can be enhanced in a targeted manner, thereby enhancing the segmentation ability of the semantic segmentation model.

[0083] See also Figure 5 , the process of double cross attention jump connection to enhance features is as follows (where ):

[0084] 21) Intermediate features of encoder output Perform block processing, that is, decompose a complete 2D image into The sequence information is expressed as:

[0085] ,

[0086] in, represents a two-dimensional average pooling operation, Reinvention, Represents a one-dimensional depthwise separable convolution.

[0087] 22) Yes Perform channel cross attention (CCA) processing and first generate The three variables, denoted as query, key, and value, are represented as follows:

[0088] , , ,

[0089] Finally, the feature map is obtained: ,

[0090] in, is a constant, CCA represents channel cross attention, refer to Activation function.

[0091] 23) Feature map Perform spatial cross attention (SCA) processing and first generate The three variables, denoted as query, key, and value, are represented as follows:

[0092] , , ;

[0093] Finally, the enhanced semantic information feature map is obtained: ,in, is a constant.

[0094] In this embodiment of the present invention, the decoder processes the value passed by the skip connection , and make predictions. The specific operations are as follows:

[0095] 31) Set mixed features ,for , ;make ,get ; IWT stands for inverse wavelet transform, which is essentially a transposed convolution that can convert frequency domain features back to spatial domain features, because the aforementioned wavelet transform (WT) converts features from the spatial domain to the frequency domain. In the decoder, the low-frequency features transmitted by the jump connection of the same layer and the mixed features transmitted by the previous layer are first added, and then the inverse wavelet transform operation is performed together with the high-frequency flux to obtain the enhanced feature information.

[0096] 32) Further, in the first layer decoder, the obtained Up-sampling, Restore to original image size and use The convolution outputs the final prediction result, which is a prediction map of the same size as the remote sensing image, which classifies the target object for each pixel in the image.

[0097] Based on the above inventive concept, the second embodiment of the present invention provides a remote sensing image semantic segmentation device based on wavelet transform convolution, which is used to implement the above remote sensing image semantic segmentation method based on wavelet transform convolution. The device includes:

[0098] A data acquisition module, used for acquiring remote sensing images to be predicted;

[0099] The remote sensing image semantic segmentation prediction model training module is used to train the remote sensing image semantic segmentation prediction model, which includes an encoder, a jump connection and a decoder; the encoder uses wavelet transform convolution, which expands the receptive field by combining wavelet transform, so that the convolutional neural network can obtain a receptive field close to the global without significantly increasing the parameters; the input is divided into different frequency bands by wavelet decomposition, allowing the convolution layer to process the low-frequency and high-frequency components separately, thereby enhancing the model's response to low-frequency components (i.e., shape features); the jump connection uses double cross attention, with the help of channel cross attention (CCA) and spatial cross attention (SCA), to enhance the effectiveness of U-Net jump connection, capture richer global context information, and enable the model to better extract fine-grained features; the decoder is used to process the decoder output features to obtain a prediction result, which is the semantic segmentation result of the remote sensing image.

[0100] The training process of the remote sensing image semantic segmentation prediction model can be found in Figure 3 , as follows:

[0101] S1. Constructing a remote sensing image semantic segmentation network model; see Figure 2 ,The remote sensing image semantic segmentation network model includes: four encoders, four skip connections and four decoders;

[0102] S2. Obtain a remote sensing image benchmark data set for labeling and divide it into a training set and a test set. Input the training set into the constructed remote sensing image semantic segmentation network model, use a coding based on wavelet transform convolution to extract the semantic feature information of the input image, use a jump connection semantic information feature map based on double cross attention, and use a decoder based on a convolutional network to output the prediction result; the label refers to the classification of each pixel in the remote sensing image. The purpose of using the real label is to compare and calculate the optimization model with the model prediction result during the model training stage; the prediction result refers to a prediction map with the same size as the input image, which classifies the target object for each pixel in the remote sensing image;

[0103] S3. Based on the loss between the prediction result and the true label, the parameters of the constructed remote sensing image semantic segmentation network model are updated, and the network model with the optimal parameters is obtained through continuous iterative training until the maximum number of iterations is reached or the model reaches the convergence condition, which is used as the remote sensing image semantic segmentation prediction model;

[0104] S4. Use the test set to test the performance of the constructed remote sensing image semantic segmentation prediction model.

[0105] The device also includes a prediction output module, which is used to input the acquired remote sensing image to be predicted into a pre-built remote sensing image semantic segmentation prediction model to obtain a semantic segmentation result of the remote sensing image.

[0106] It should be noted that the semantic segmentation result of remote sensing images refers to: the semantic segmentation of remote sensing images performs pixel-level image classification on remote sensing images. The semantic segmentation result of remote sensing images is a prediction map of the same size as the remote sensing image, which classifies the target object for each pixel in the image. The prediction map labels the target object pixels, and different colors represent different types of objects. For example, yellow represents cars, green represents trees, blue represents buildings, and red pixels represent backgrounds.

[0107] The specific working process of each module described above can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here. The module division is only a logical function division, and there may be other division methods in actual implementation, such as multiple modules can be combined or integrated into another system.

[0108] Based on the same inventive concept, the present invention provides a computer system, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the computer program is loaded into the processor, the steps of the remote sensing image semantic segmentation method based on wavelet transform convolution are implemented.

[0109] Based on the same inventive concept, the present invention provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the steps of the above-mentioned remote sensing image semantic segmentation method based on wavelet transform convolution are implemented.

[0110] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0111] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes 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.

[0112] 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.

[0113] 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 A step that specifies a function in one or more boxes.

[0114] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A remote sensing image semantic segmentation method based on wavelet transform convolution, characterized in that: include: Acquire remote sensing images to be predicted; Inputting the remote sensing image to be predicted into a pre-built remote sensing image semantic segmentation prediction model to obtain a semantic segmentation result of the remote sensing image; the semantic segmentation result of the remote sensing image refers to the classification of target objects for each pixel in the remote sensing image; The remote sensing image semantic segmentation prediction model is constructed as follows: Construct a remote sensing image semantic segmentation network model; The remote sensing image semantic segmentation network model includes: an encoder, a skip connection and a decoder; the encoder is used to extract a semantic information feature map of an input image, the skip connection is used to enhance the encoder output features, and the decoder is used to output a prediction result; the prediction result refers to a prediction map with the same size as the input image, and the prediction map classifies the target object for each pixel in the remote sensing image; A remote sensing image benchmark data set is obtained for label annotation and divided into a training set and a test set. The remote sensing image semantic segmentation network model is trained based on the training set to obtain a network model with optimal parameters as a remote sensing image semantic segmentation prediction model.

2. The remote sensing image semantic segmentation method based on wavelet transform convolution according to claim 1, characterized in that: Before each iterative training, the images in the training set and the test set are randomly cropped, randomly horizontally slid, randomly flipped, rotated, and color enhanced.

3. The remote sensing image semantic segmentation method based on wavelet transform convolution according to claim 1, characterized in that: The remote sensing image semantic segmentation network model includes: four encoders, four skip connections and four decoders; The encoder extracts a semantic information feature map of the input image based on wavelet transform convolution; The skip connection is based on channel cross attention and spatial cross attention to enhance the semantic information feature map.

4. The remote sensing image semantic segmentation method based on wavelet transform convolution according to claim 3 is characterized in that: The encoder extracts a semantic information feature map of an input image based on wavelet transform convolution, including: Input remote sensing image use Convolution extracts features and obtains feature maps ; For feature maps Using wavelet decomposition, let the initial low-pass component , intermediate features ,in is the initial weight, and four fluxes of different frequencies are obtained by cascade wavelet decomposition: , , which correspond to the semantic information feature maps obtained by encoders at different layers, and convolution operations are performed on fluxes of different frequencies to obtain four intermediate features of different frequencies: ;in, Refers to wavelet transform, is the convolution operation, is the weight.

5. The remote sensing image semantic segmentation method based on wavelet transform convolution according to claim 4 is characterized in that: The jump connection is based on channel cross attention and spatial cross attention to enhance the semantic information feature map, including: Intermediate features of the encoder output After block processing, we get , : ， in, represents a two-dimensional average pooling operation, Reinvention, Represents a one-dimensional depth-wise separable convolution; right Perform channel cross attention processing and first generate Three variables, denoted as query, key, and value, are represented as: ， ， ， Finally, the feature map is obtained: , in, is a constant, represents channel cross attention, refer to Activation function; For feature maps Perform spatial cross attention processing and first generate Three variables, denoted as query, key, and value, are represented as: ， ， ； Finally, the enhanced semantic information feature map is obtained: ,in, is a constant.

6. The remote sensing image semantic segmentation method based on wavelet transform convolution according to claim 5, characterized in that: The decoder outputs a prediction result, including: Set mixed features ,for , ; make ,get ; Where IWT stands for inverse wavelet transform, In the first decoder, the obtained Up-sampling, Restore to original image size and use The convolution outputs the final prediction result.

7. The remote sensing image semantic segmentation method based on wavelet transform convolution according to claim 1, characterized in that: During the training process of the remote sensing image semantic segmentation network model, a cross entropy loss function is used to calculate the loss, and the model parameters are updated by back propagation.

8. A remote sensing image semantic segmentation device based on wavelet transform convolution, characterized in that: The device for implementing the remote sensing image semantic segmentation method based on wavelet transform convolution according to any one of claims 1 to 6 comprises: A data acquisition module, used for acquiring remote sensing images to be predicted; Remote sensing image semantic segmentation prediction model training module, used to train remote sensing image semantic segmentation prediction model; A prediction output module, used for inputting the remote sensing image to be predicted into the remote sensing image semantic segmentation prediction model to obtain a semantic segmentation result of the remote sensing image; the remote sensing semantic segmentation result refers to the identification and classification of ground objects for each pixel in the remote sensing image; The training process of the remote sensing image semantic segmentation prediction model training module is as follows: Constructing a remote sensing image semantic segmentation network model; the remote sensing image semantic segmentation network model comprises: an encoder, a skip connection and a decoder; the encoder is used to extract a semantic information feature map of an input image, the skip connection is used to enhance the encoder output feature, and the decoder is used to output a prediction result; the prediction result refers to a prediction map with the same size as the input image, and the prediction map classifies the target object for each pixel in the remote sensing image; A remote sensing image benchmark data set is obtained for label annotation and divided into a training set and a test set. The remote sensing image semantic segmentation network model is trained based on the training set to obtain a network model with optimal parameters as a remote sensing image semantic segmentation prediction model.

9. A computer-readable storage medium storing one or more programs, characterized in that: The one or more programs include instructions, which, when executed by a computing device, cause the computing device to perform any one of the remote sensing image semantic segmentation methods based on wavelet transform convolution according to claims 1 to 6.

10. A computing device, characterized in that include, One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include instructions for executing any one of the remote sensing image semantic segmentation methods based on wavelet transform convolution according to claims 1 to 6.

Images