Remote Sensing Image Super-Resolution Reconstruction Method Based on Degradation Mechanism

Abstract

The present invention belongs to the technical field of image processing, and particularly relates to a super-resolution reconstruction method for remote sensing images based on a degradation mechanism. It includes: S1: Construct a degradation model, input a high-resolution remote sensing image into the degradation model, and obtain a degraded image corresponding to the high-resolution remote sensing image; S2: Construct a super-resolution network, input the degraded image into the super-resolution network, and obtain a reconstructed image; S3: Use a joint loss function to train the super-resolution network to obtain a trained super-resolution network; S4: Input the degraded image to be reconstructed into the trained super-resolution network to obtain a reconstructed super-resolution image. The present invention not only improves the resolution of remote sensing images, but also significantly improves the quality and processing efficiency of the images, and has high practical application value.

Description

Technical Field

[0001] The present invention belongs to the technical field of image processing, and particularly relates to a super-resolution reconstruction method for remote sensing images based on a degradation mechanism. Background Art

[0002] The super-resolution reconstruction technology of remote sensing images has received extensive attention in recent years. Its main purpose is to improve the detail performance of remote sensing images by increasing the image resolution and enhance the accuracy of target detection, classification, and recognition. Currently, various methods have been applied to the super-resolution reconstruction of remote sensing images, mainly including interpolation methods, reconstruction methods based on signal processing, and deep learning methods.

[0003] The interpolation method is a traditional super-resolution reconstruction method. Typical techniques include bilinear interpolation and bicubic interpolation, etc. These interpolation methods fill out the high-resolution image by performing interpolation calculations between the pixels of the original low-resolution image. This method is simple and has high computational efficiency. However, since it cannot effectively restore the high-frequency details of the image, the reconstructed result has the disadvantages of blurriness and unclearness, especially when dealing with complex textures or edge information, the effect is not good.

[0004] The super-resolution reconstruction method based on signal processing utilizes the features of the image in the spatial domain or frequency domain and realizes the super-resolution reconstruction of the degraded image by solving specific optimization problems. In this type of reconstruction method, some studies use regularization techniques to constrain the reconstruction result of the image, thereby reducing the influence of noise and artifacts. However, when the computational complexity is high and the processed image is a real-world remote sensing image, the super-resolution reconstruction method based on signal processing often has limited reconstruction quality due to the lack of high-frequency information in the image.

[0005] In recent years, deep learning technology has made remarkable progress in the field of image super-resolution reconstruction. In particular, deep models such as convolutional neural networks (CNNs) have been widely applied to the super-resolution reconstruction of remote sensing images due to their powerful feature extraction and learning capabilities. The super-resolution method based on deep learning usually trains a deep neural network and uses paired high-resolution and low-resolution images for supervised learning to obtain higher-quality reconstructed images. However, most of the existing deep learning methods are based on idealized degradation models, which do not fully match the complex degradation mechanism in the actual remote sensing imaging process, resulting in the quality of the reconstructed result still needing to be improved when dealing with real remote sensing images.

[0006] Existing super-resolution reconstruction methods have their own advantages and disadvantages. Traditional interpolation methods are computationally simple but lack the ability to recover high-frequency details. Although reconstruction methods based on signal processing can improve the reconstruction effect through mathematical models, they are computationally complex and perform poorly when dealing with real remote sensing images. Deep learning methods perform well in terms of reconstruction quality, but most current methods rely on idealized degradation models and are difficult to accurately simulate the complex degradation process of remote sensing imaging. There are significant differences between the existing degradation models and the actual remote sensing imaging process, resulting in low-quality reconstructed images that cannot fully restore the high-frequency details and texture information in the images, affecting their performance in practical applications.

[0007] Most existing methods are unable to effectively handle multiple degradation factors in the complex imaging chain when processing remote sensing images, such as atmospheric scattering, blur in the optical system, and noise interference. Although deep learning-based super-resolution methods have improved the reconstruction effect to a certain extent, the design of their network structure and loss function is still imperfect, making it difficult to balance the extraction and utilization of local and global feature information, resulting in insufficient visual effects and information fidelity of the reconstructed images. Summary of the Invention

[0008] In view of this, the present invention aims to provide a super-resolution reconstruction method for remote sensing images based on a degradation mechanism to solve the problems that there are significant differences between the degradation model in existing super-resolution reconstruction methods and the actual remote sensing imaging process, resulting in low-quality reconstructed images that cannot fully restore the high-frequency details and texture information in the images, and most existing methods are unable to effectively handle multiple degradation factors in the complex imaging chain, such as atmospheric scattering, blur in the optical system, and noise interference. The present invention significantly improves the quality and processing efficiency of images while enhancing the resolution of remote sensing images, and has high practical application value.

[0009] To achieve the above object, the technical solution of the present invention is realized as follows:

[0010] A super-resolution reconstruction method for remote sensing images based on a degradation mechanism specifically includes the following steps:

[0011] S1: Construct a degradation model, input a high-resolution remote sensing image into the degradation model, and obtain a degraded image corresponding to the high-resolution remote sensing image;

[0012] S2: Construct a super-resolution network, input the degraded image into the super-resolution network, and obtain a reconstructed image;

[0013] The super-resolution network includes a spatial domain branch and a frequency domain branch. The spatial domain branch extracts local multi-scale feature information of the degraded image through multiple levels of convolutional operations to obtain spatial domain features; the frequency domain branch analyzes and processes the global frequency features of the degraded image to obtain frequency domain features; the spatial domain features and frequency domain features are fused to obtain a reconstructed image;

[0014] S3: Use the joint loss function to train the super-resolution network to obtain a trained super-resolution network;

[0015] The joint loss function includes a pixel-level loss function, a multi-scale feature loss function, and a global feature loss function;

[0016] S4: Input the degraded image to be reconstructed into the trained super-resolution network to obtain a reconstructed super-resolution image.

[0017] Further, in step S1, the degradation model includes a spatial domain filtering module, a frequency domain filtering module, and a degradation factor random control module. Input the high-resolution remote sensing image into the spatial domain filtering module for processing to obtain a first spatial domain image; after performing Fourier transform on the high-resolution remote sensing image, input it into the frequency domain filtering module. The degradation factor random control module controls the frequency domain filtering module to randomly introduce degradation factors to the high-resolution remote sensing image after Fourier transform, and perform inverse Fourier transform on the high-resolution remote sensing image with the introduced degradation factors to obtain a second spatial domain image; synthesize the first spatial domain image and the second spatial domain image and then perform downsampling operation and noise introduction operation to obtain a degraded image.

[0018] Further, the frequency domain filtering module includes an atmospheric disturbance factor, an optical system disturbance factor, a first platform component disturbance factor, a second platform component disturbance factor, and a third platform component disturbance factor;

[0019] The atmospheric disturbance factor Ⅰ is: ;

[0020] ;

[0021] Among them, is the angular spatial frequency, , is the focal length of the optical system, is the spatial frequency, is the refractive index structure constant, is the wavelength, R is the propagation path, a1 is the turbulence effect, a2 is the aerosol effect, is the modulation transfer function of the turbulence effect, is the modulation transfer function of the aerosol effect;

[0022] ;

[0023] wherein, is the scattering coefficient of atmospheric aerosol, is the absorption coefficient of atmospheric aerosol, is the angular spatial cut-off frequency of the aerosol;

[0024] The optical system perturbation factor II is: ;

[0025] ;

[0026] wherein, is , is the spatial frequency in the horizontal direction, is the spatial cut-off frequency, = , D is the diameter of the optical system, b1 is the diffraction effect, b2 is the defocus effect, b3 is the image shift effect, is the modulation transfer function of the diffraction effect, is the modulation transfer function of the defocus effect, is the modulation transfer function of the image shift effect;

[0027] ;

[0028] wherein, is the first-order Bessel function, , is the defocus amount, F is the focal ratio of the optical system, is an intermediate parameter without physical meaning;

[0029] ;

[0030] ;

[0031] wherein, is the image shift generated by the change of satellite velocity, is the image shift generated by satellite yaw, is the image shift generated by the change of satellite attitude angle, is the image shift generated by the earth's rotation, is the image shift generated by the random vibration of the satellite and the optical remote sensor, is the image shift generated by the jitter of the integration time, is an intermediate variable without physical meaning;

[0032] The first platform component perturbation factor III is: ;

[0033] ;

[0034] where N is the sampling frequency and d is the diameter of the dispersion spot, is low-frequency vibration, is high-frequency vibration;

[0035] ;

[0036] where is the zero-order Bessel function, is the platform amplitude, is the spatial frequency;

[0037] The perturbation factor Ⅳ of the second platform component is:

[0038] ;

[0039] ;

[0040] where is the spatial frequency along the scanning direction, is the CCD pixel size, d1 is the spatial response, d2 is the time response, d3 is the CCD response, is the modulation transfer function of the spatial response, is the modulation transfer function of the time response, is the modulation transfer function of the CCD response;

[0041] ;

[0042] where is the time frequency value along the scanning direction, is the 3-dB point frequency value of the detector time response;

[0043] ;

[0044] where is the CCD transfer charge amount, is the charge transfer efficiency, is the sampling frequency of the CCD signal;

[0045] The perturbation factor Ⅴ of the third platform component is:

[0046] ;

[0047] ;

[0048] where is the frequency value of the -3dB point of the preamplifier response, e1 is the preamplifier response, e2 is the postamplifier response, e3 is the lift response, is the modulation transfer function of the preamplifier response, is the modulation transfer function of the post-amplifier response, is the modulation transfer function of the lift response;

[0049] ;

[0050] in, The frequency value of the -3dB point of the post-amplifier response;

[0051] ;

[0052] in, is the frequency value when the boost pressure is maximum, for The magnitude of the value at .

[0053] Furthermore, the formula for the degradation factor random control module to control the frequency domain filtering module to randomly introduce degradation factors into the high-resolution remote sensing image after Fourier transformation is:

[0054] K= · · ;

[0055] Among them, U is uniform distribution, X1~X5 are coefficients that obey uniform distribution, n1 is 0, and n2 is 1.

[0056] Further, in step S2, the spatial domain branch includes a first 3×3 unconstrained blueprint convolution layer and n cascaded basic units, each basic unit includes a cascaded first basic module and a second basic module, and the output of the first basic module is the input of the second basic module;

[0057] The degraded image is copied four times, and three of the degraded images are subjected to RGB three-channel extraction operations, and the remaining degraded image is subjected to Y channel extraction operations, and all the extracted feature maps are subjected to feature splicing operations to obtain a spliced ​​feature map, and the spliced ​​feature map is subjected to a convolution operation by the first 3×3 unconstrained blueprint convolution layer to obtain an intermediate feature map, and the intermediate feature map is input into n cascaded basic units for processing to obtain a first feature map;

[0058] The frequency domain branch includes a mask, a second 3×3 unconstrained blueprint convolution layer, a third 3×3 unconstrained blueprint convolution layer, a fourth 3×3 unconstrained blueprint convolution layer, a first 3×3 convolution layer, a second 3×3 convolution layer, a third 3×3 convolution layer, a first batch of normalization layers, a second batch of normalization layers, a third batch of normalization layers, a first SiLU activation function, a second SiLU activation function, and a third SiLU activation function;

[0059] Perform Fourier transform on the degraded image to obtain a frequency-domain image, perform a dot product operation on the frequency-domain image and the mask to obtain frequency-domain images under different masks, and after processing the frequency-domain images under different masks through an attention mechanism, obtain a second feature map, a third feature map, and a fourth feature map;

[0060] Input the second feature map into a second 3×3 unconstrained blueprint convolutional layer for processing to obtain a fifth feature map. After sequentially processing the fifth feature map through a first 3×3 convolutional layer, a first batch normalization layer, and a first SiLU activation function, obtain a sixth feature map. Concatenate the fifth feature map and the sixth feature map to obtain a seventh feature map, and perform inverse Fourier transform on the seventh feature map to obtain an eighth feature map; Input the third feature map into a third 3×3 unconstrained blueprint convolutional layer for processing to obtain a ninth feature map. After sequentially processing the ninth feature map through a second 3×3 convolutional layer, a second batch normalization layer, and a second SiLU activation function, obtain a tenth feature map. Concatenate the ninth feature map and the tenth feature map to obtain an eleventh feature map, and perform inverse Fourier transform on the eleventh feature map to obtain a twelfth feature map; Input the fourth feature map into a fourth 3×3 unconstrained blueprint convolutional layer for processing to obtain a thirteenth feature map. After sequentially processing the thirteenth feature map through a third 3×3 convolutional layer, a third batch normalization layer, and a third SiLU activation function, obtain a fourteenth feature map. Concatenate the thirteenth feature map and the fourteenth feature map to obtain a fifteenth feature map, and perform inverse Fourier transform on the fifteenth feature map to obtain a sixteenth feature map.

[0061] Furthermore, concatenate the output features of the first basic module of each basic unit, the first feature map, the eighth feature map, the twelfth feature map, and the sixteenth feature map to obtain a seventeenth feature map. After sequentially performing convolutional processing on the seventeenth feature map through a first 1×1 convolutional layer and a fifth 3×3 unconstrained blueprint convolutional layer, obtain an eighteenth feature map with supplemented features. After adding the eighteenth feature map and the intermediate feature map, obtain a nineteenth feature map. The nineteenth feature map is sequentially processed through a fourth 3×3 convolutional layer and upsampling to obtain a reconstructed image.

[0062] Furthermore, in step S3, the output of the joint loss function is the sum of the pixel-level loss function, the multi-scale feature loss function, and the global feature loss function;

[0063] The pixel-level loss function is the L1 loss function, and the input of the pixel-level loss function is the degraded image and the reconstructed image;

[0064] The multi-scale feature loss function is the Huber loss function with a parameter of 1, and the input of the multi-scale feature loss function is the reconstructed concatenated image and the high-resolution concatenated image;

[0065] The global feature loss function is the Huber loss function with a parameter of 1.25, and the input of the global feature loss function is the reconstructed stitched image and the high-resolution stitched image.

[0066] Furthermore, the way to obtain the reconstructed stitched image is as follows: the reconstructed images are respectively input into the global feature extraction module and the local feature extraction module for feature extraction, and the extracted global features and all local features are stitched together to obtain the reconstructed stitched image;

[0067] The way to obtain the high-resolution stitched image is as follows: the high-resolution remote sensing images are respectively input into the global feature extraction module and the local feature extraction module for feature extraction, and the extracted global features and all local features are stitched together to obtain the high-resolution stitched image.

[0068] Furthermore, the global feature extraction module includes a fifth 3×3 convolutional layer, an activation layer sub-module, a pooling layer sub-module, a fully connected layer sub-module, a flattening and transposing sub-module, a first encoder, a second encoder, a third encoder, a fourth encoder, and a fifth encoder;

[0069] The degraded image or the reconstructed image is input into the global feature extraction module. The degraded image or the reconstructed image is sequentially processed by the fifth 3×3 convolutional layer, the activation layer sub-module, the pooling layer sub-module, and the fully connected layer sub-module to obtain the feature map A1. After the feature map A1 is flattened and transposed by the flattening and transposing sub-module, the feature map A2 is obtained. The feature map A2 is input into the first encoder for encoding to obtain the feature map A3. The feature map A3 is input into the second encoder for encoding to obtain the feature map A4. The feature map A4 is input into the third encoder for encoding to obtain the feature map A5. The feature map A5 is input into the fourth encoder for encoding to obtain the feature map A6. The feature map A6 is input into the fifth encoder for encoding to obtain the feature map A7. The attention mechanism is used to process the feature maps A3, A4, A5, A6, and A7 to correspondingly obtain the global features of the degraded image or the reconstructed image.

[0070] Furthermore, the local feature extraction module includes a first basic unit convolutional layer, a second basic unit convolutional layer, a third basic unit convolutional layer, a fourth basic unit convolutional layer, a fifth basic unit convolutional layer, a first activation layer, a second activation layer, a third activation layer, a fourth activation layer, a fifth activation layer, a first max pooling layer, a second max pooling layer, a third max pooling layer, a fourth max pooling layer, and a fifth max pooling layer;

[0071] Input the degraded image or the reconstructed image into the local feature extraction module. The degraded image or the reconstructed image is successively processed by the first basic unit convolutional layer, the first activation layer, and the first max pooling layer to obtain the feature map B1. The feature map B1 is successively processed by the second basic unit convolutional layer, the second activation layer, and the second max pooling layer to obtain the feature map B2. The feature map B2 is successively processed by the third basic unit convolutional layer, the third activation layer, and the third max pooling layer to obtain the feature map B3. The feature map B3 is successively processed by the fourth basic unit convolutional layer, the fourth activation layer, and the fourth max pooling layer to obtain the feature map B4. The feature map B4 is successively processed by the fifth basic unit convolutional layer, the fifth activation layer, and the fifth max pooling layer to obtain the feature map B5. The attention mechanism is used to process the feature maps B1, B2, B3, B4, and B5 to correspondingly obtain the local features of the degraded image or the reconstructed image.

[0072] Compared with the prior art, the present invention can achieve the following beneficial effects:

[0073] (1) For the method for super-resolution reconstruction of remote sensing images based on a degradation mechanism according to the present invention, through the degradation model based on the degradation mechanism and the super-resolution network in the joint spatial-frequency domain, it can better simulate the degradation process in remote sensing imaging, effectively improve the quality of the super-resolution image, especially outstanding in the aspects of high-frequency detail and texture restoration, and significantly improve the quality of the reconstructed image.

[0074] (2) For the method for super-resolution reconstruction of remote sensing images based on a degradation mechanism according to the present invention, the degradation space is expanded through the degradation factor random control module, making the method for super-resolution reconstruction of remote sensing images of the present invention have stronger adaptability and robustness, capable of processing different types of remote sensing images, and showing better generalization ability in practical applications.

[0075] (3) For the method for super-resolution reconstruction of remote sensing images based on a degradation mechanism according to the present invention, the super-resolution network in the joint spatial domain and frequency domain enables the present invention to fully extract and utilize the local and global feature information of the degraded image, thereby improving the fineness and overall visual effect of the reconstructed image.

[0076] (4) For the method for super-resolution reconstruction of remote sensing images based on a degradation mechanism according to the present invention, through the design of the joint loss function, the reconstructed image is optimized simultaneously at three levels of pixel level, global features, and multi-scale features, ensuring high-quality super-resolution output.

[0077] (5)The super-resolution reconstruction method of remote sensing images based on the degradation mechanism according to the present invention can be applied to fields such as satellite remote sensing, environmental monitoring, urban planning, and national defense security. By improving the detail restoration ability of degraded images, the spatial resolution and quality of the reconstructed images are enhanced. Specific applications include, but are not limited to, the construction of high-precision geographic information systems (GIS), remote sensing image classification, target detection and recognition, etc. Specifically: 1) Satellite remote sensing: Improve the resolution of remote sensing images and enhance the recognition and monitoring capabilities of ground targets. 2) Environmental monitoring: Used to track changes in dynamic environments such as pollution diffusion and natural disasters. 3) Urban planning: Provide more accurate basic data by enhancing image resolution to support building and road planning. 4) Agricultural and forestry monitoring: Used for refined management of agriculture and forestry such as crop health assessment and forest cover monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0078] The accompanying drawings that form a part of the present invention are used to provide a further understanding of the present invention. The schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0079] Figure 1 It is a schematic flow chart of the super-resolution reconstruction method of remote sensing images based on the degradation mechanism according to the embodiment of the present invention;

[0080] Figure 2 It is a structural block diagram of the super-resolution reconstruction method of remote sensing images based on the degradation mechanism according to the embodiment of the present invention;

[0081] Figure 3 It is a schematic network structure diagram of the degradation model according to the embodiment of the present invention;

[0082] Figure 4 It is a schematic network structure diagram of the super-resolution network according to the embodiment of the present invention;

[0083] Figure 5 It is a schematic network structure diagram of the global feature extraction module according to the embodiment of the present invention;

[0084] Figure 6 It is a schematic network structure diagram of the local feature extraction module according to the embodiment of the present invention;

[0085] Figure 7 It is a schematic network structure diagram of the first basic module according to the embodiment of the present invention;

[0086] Figure 8 It is a schematic network structure diagram of the second basic module according to the embodiment of the present invention;

[0087] Figure 9The reconstruction effect diagram for reconstructing a severely degraded remote sensing image according to an embodiment of the present invention;

[0088] Figure 10 The reconstruction effect diagram for reconstructing a first type of remote sensing image with complex texture according to an embodiment of the present invention;

[0089] Figure 11 The reconstruction effect diagram for reconstructing a second type of remote sensing image with complex texture according to an embodiment of the present invention. Detailed implementation manners

[0090] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and do not constitute a limitation to the present invention.

[0091] It should be noted that, without conflict, the embodiments in the present invention and the features in the embodiments may be combined with each other.

[0092] In the description of the present invention, it should be understood that the orientation or positional relationships indicated by the terms "center", "longitudinal", "transverse", "up", "down", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. are based on the orientation or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as a limitation to the present invention. In addition, the terms "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, the features defined with "first", "second", etc. may explicitly or implicitly include one or more of such features. In the description of the present invention, unless otherwise specified, the meaning of "a plurality" is two or more.

[0093] In the description of the present invention, it should be noted that, unless otherwise clearly defined and limited, the terms "installed", "connected", "connected" should be understood in a broad sense. For example, it may be a fixed connection, a detachable connection, or an integral connection; it may be a mechanical connection or an electrical connection; it may be directly connected or indirectly connected through an intermediate medium, and it may be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood through specific situations.

[0094] The present invention will be described in detail below with reference to the drawings and in combination with embodiments.

[0095] As shown Figures 1 to 2 in the figure, a super-resolution reconstruction method for remote sensing images based on a degradation mechanism specifically includes the following steps:

[0096] S1: Construct a degradation model, input a high-resolution remote sensing image into the degradation model, and obtain a degraded image corresponding to the high-resolution remote sensing image;

[0097] S2: Construct a super-resolution network, input the degraded image into the super-resolution network, and obtain a reconstructed image;

[0098] The super-resolution network includes a spatial domain branch and a frequency domain branch. The spatial domain branch extracts local multi-scale feature information of the degraded image through multiple levels of convolution operations to obtain spatial domain features; the frequency domain branch analyzes and processes the global frequency features of the degraded image to obtain frequency domain features; the spatial domain features and frequency domain features are fused to obtain a reconstructed image;

[0099] S3: Use a joint loss function to train the super-resolution network to obtain a trained super-resolution network;

[0100] The joint loss function includes a pixel-level loss function, a multi-scale feature loss function, and a global feature loss function;

[0101] S4: Input the degraded image to be reconstructed into the trained super-resolution network to obtain a reconstructed super-resolution image.

[0102] The present invention designs a degradation model based on the remote sensing imaging chain. This model simulates various degradation factors that may exist in the remote sensing imaging process, including atmospheric disturbance, optical system blur, sensor noise, etc. By introducing a degradation factor random control module, the expression ability of the degradation space is expanded, making the model more generalizable and robust, and ensuring that the model can adapt to different degradation situations.

[0103] The network structure of the degradation model is specifically introduced below. As shown Figure 3 in the figure, the degradation model includes a spatial domain filtering module, a frequency domain filtering module, and a degradation factor random control module. Input a high-resolution remote sensing image into the spatial domain filtering module for processing to obtain a first spatial domain image; input the high-resolution remote sensing image after Fourier transform (FFT) into the frequency domain filtering module. The degradation factor random control module controls the frequency domain filtering module to randomly introduce degradation factors to the high-resolution remote sensing image after Fourier transform, and perform inverse Fourier transform (iFFT) on the high-resolution remote sensing image with the introduced degradation factors to obtain a second spatial domain image; synthesize the first spatial domain image and the second spatial domain image and then perform downsampling operation and noise introduction operation to obtain a degraded image. This degradation model not only enhances the adaptability to complex degradation scenarios but also effectively improves the accuracy of super-resolution reconstruction.

[0104] The frequency domain filtering module includes atmospheric disturbance factors, optical system disturbance factors, first platform component disturbance factors, second platform component disturbance factors, and third platform component disturbance factors (the first platform component disturbance factors, second platform component disturbance factors, and third platform component disturbance factors are sensor noise factors).

[0105] The atmospheric disturbance factor Ⅰ is: ;

[0106] ;

[0107] Wherein, is the angular spatial frequency, , is the focal length of the optical system, is the spatial frequency, is the refractive index structure constant, is the wavelength, R is the propagation path, a1 is the turbulence effect, a2 is the aerosol effect, is the modulation transfer function of the turbulence effect, is the modulation transfer function of the aerosol effect;

[0108] ;

[0109] Wherein, is the scattering coefficient of atmospheric aerosol, is the absorption coefficient of atmospheric aerosol, is the angular spatial cut-off frequency of the aerosol;

[0110] The optical system disturbance factor Ⅱ is: ;

[0111] ;

[0112] Wherein, is , is the spatial frequency in the horizontal direction, is the spatial cut-off frequency, = , D is the diameter of the optical system, b1 is the diffraction effect, b2 is the defocus effect, b3 is the image shift effect, is the modulation transfer function of the diffraction effect, is the modulation transfer function of the defocus effect, is the modulation transfer function of the image shift effect;

[0113] ;

[0114] Wherein, is the first-order Bessel function, , where is the defocus amount and F is the focal ratio of the optical system, is an intermediate parameter without physical meaning;

[0115] ;

[0116] ;

[0117] Among them, is the image motion caused by the change in satellite velocity, is the image motion caused by satellite yaw, is the image motion caused by the change in satellite attitude angle, is the image motion caused by the earth's rotation, is the image motion caused by the random vibration of the satellite and the optical remote sensor, is the image motion caused by the jitter of the integration time, is an intermediate variable without physical meaning;

[0118] The first platform component disturbance factor Ⅲ is: ;

[0119] ;

[0120] Among them, N is the sampling frequency, d is the diameter of the blur spot, is the low-frequency vibration, is the high-frequency vibration;

[0121] ;

[0122] Among them, is the zero-order Bessel function, is the platform amplitude, is the spatial frequency;

[0123] The second platform component disturbance factor Ⅳ is:

[0124] ;

[0125] ;

[0126] Among them, is the spatial frequency along the scanning direction, is the CCD pixel size, d1 is the spatial response, d2 is the time response, d3 is the CCD response, is the modulation transfer function of the spatial response, is the modulation transfer function of the time response, is the modulation transfer function of the CCD response;

[0127] ;

[0128] Among them, is the time - frequency value along the scanning direction, is the 3 - dB point frequency value of the detector time response;

[0129] ;

[0130] Among them, is the CCD transfer charge amount, is the charge transfer efficiency, is the sampling frequency of the CCD signal;

[0131] The disturbance factor Ⅴ of the third platform component is:

[0132] ;

[0133] ;

[0134] Among them, is the frequency value at the - 3dB point of the pre - amplifier response, e1 is the pre - amplifier response, e2 is the post - amplifier response, e3 is the lifting response, is the modulation transfer function of the pre - amplifier response, is the modulation transfer function of the post - amplifier response, is the modulation transfer function of the lifting response;

[0135] ;

[0136] Among them, is the frequency value at the - 3dB point of the post - amplifier response;

[0137] ;

[0138] Among them, is the frequency value when the supercharging is maximum, is the amplitude value at;

[0139] The formula for the degradation factor random control module to control the frequency - domain filtering module to randomly introduce degradation factors into the high - resolution remote - sensing image after Fourier transform is:

[0140] K = · · ;

[0141] Among them, U is a uniform distribution, X1~X5 are coefficients obeying the uniform distribution, n1 is 0, and n2 is 1.

[0142] In the design of the super-resolution network, the present invention proposes a super-resolution network that combines the spatial domain and the frequency domain (JSF-SRNet). Among them, the spatial domain branch is responsible for extracting the local multi-scale feature information of the degraded image, and the frequency domain branch is used to capture the global frequency features of the degraded image. The two branches achieve information complementarity between the spatial domain and the frequency domain through a specific fusion mechanism, making full use of the multi-dimensional feature information of the image to improve the reconstruction effect. Specifically, the spatial domain branch processes the texture and details of the degraded image at different scales, and the frequency domain branch focuses on the restoration of high-frequency information through frequency separation technology, ensuring the clarity and structural integrity of the reconstructed image.

[0143] As Figure 4 shown, the spatial domain branch includes a first 3×3 unconstrained blueprint convolutional layer and n cascaded basic units. Each basic unit includes a cascaded first basic module and a second basic module, and the output of the first basic module is the input of the second basic module;

[0144] Copy the degraded image four times, perform RGB three-channel extraction operations on three of the degraded images, perform Y-channel extraction operations on the remaining one degraded image, perform feature stitching operations on all the extracted feature maps to obtain a stitched feature map, and perform convolution operations on the stitched feature map through the first 3×3 unconstrained blueprint convolutional layer (the unconstrained blueprint convolutional layer is BSConv-U (Unconstrained Blueprint Convolution, unconstrained blueprint separable convolution), which has the advantages of few parameters (lightweight), fast operation speed, and enhanced intra-kernel correlation of the convolutional kernel) to obtain an intermediate feature map, and input the intermediate feature map into n cascaded basic units for processing to obtain a first feature map.

[0145] The frequency domain branch includes a mask (specifically a low-pass filter, a high-pass filter, or a band-pass filter), a second 3×3 unconstrained blueprint convolutional layer, a third 3×3 unconstrained blueprint convolutional layer, a fourth 3×3 unconstrained blueprint convolutional layer, a first 3×3 convolutional layer, a second 3×3 convolutional layer, a third 3×3 convolutional layer, a first batch normalization layer, a second batch normalization layer, a third batch normalization layer, a first SiLU activation function, a second SiLU activation function, and a third SiLU activation function;

[0146] Perform a fast Fourier transform on the degraded image ( Figure 4 denoted by F in

[0147] The second feature map is input into the second 3×3 unconstrained blueprint convolutional layer for processing to obtain the fifth feature map. After the fifth feature map is successively processed by the first 3×3 convolutional layer, the first batch normalization layer, and the first SiLU activation function, the sixth feature map is obtained. The fifth feature map and the sixth feature map are feature concatenated ( Figure 4 denoted by C in Figure 4 ) to obtain the seventh feature map. The seventh feature map is subjected to an inverse fast Fourier transform ( denoted by iF in

[0148] ) to obtain the eighth feature map. The third feature map is input into the third 3×3 unconstrained blueprint convolutional layer for processing to obtain the ninth feature map. After the ninth feature map is successively processed by the second 3×3 convolutional layer, the second batch normalization layer, and the second SiLU activation function, the tenth feature map is obtained. The ninth feature map and the tenth feature map are feature concatenated to obtain the eleventh feature map. The eleventh feature map is subjected to an inverse Fourier transform to obtain the twelfth feature map. The fourth feature map is input into the fourth 3×3 unconstrained blueprint convolutional layer for processing to obtain the thirteenth feature map. After the thirteenth feature map is successively processed by the third 3×3 convolutional layer, the third batch normalization layer, and the third SiLU activation function, the fourteenth feature map is obtained. The thirteenth feature map and the fourteenth feature map are feature concatenated to obtain the fifteenth feature map. The fifteenth feature map is subjected to an inverse Fourier transform to obtain the sixteenth feature map. Figure 4

[0149] The present invention also proposes a joint loss function (JLF-PMG). This loss function optimizes the network from three dimensions: pixel level, global feature level, and multi-scale feature level to ensure that the super-resolution image has higher visual quality and structural integrity while retaining details. The pixel-level loss function calculates the pixel difference between the reconstructed image and the high-resolution remote sensing image through the L1 norm. The multi-scale feature loss function extracts local features of different scales through a convolutional neural network and calculates the difference between the reconstructed image features and the high-resolution remote sensing image features. The global feature loss function extracts the global context information of the degraded image through a transformation network (such as a ViT model) to ensure the consistency of the overall structure of the degraded image.

[0150] The output of the combined loss function is the sum of the pixel-level loss function, the multi-scale feature loss function, and the global feature loss function;

[0151] The pixel-level loss function is the L1 loss function, and the input of the pixel-level loss function is the degraded image and the reconstructed image;

[0152] The multi-scale feature loss function is the Huber loss function with a parameter of 1, and the input of the multi-scale feature loss function is the reconstructed mosaic image and the high-resolution mosaic image;

[0153] The global feature loss function is the Huber loss function with a parameter of 1.25, and the input of the global feature loss function is the reconstructed mosaic image and the high-resolution mosaic image.

[0154] The way to obtain the reconstructed mosaic image is as follows: the reconstructed image is respectively input into the global feature extraction module and the local feature extraction module for feature extraction, and the extracted global features and all local features are mosaicked to obtain the reconstructed mosaic image;

[0155] The way to obtain the high-resolution mosaic image is as follows: the high-resolution remote sensing image is respectively input into the global feature extraction module and the local feature extraction module for feature extraction, and the extracted global features and all local features are mosaicked to obtain the high-resolution mosaic image.

[0156] As Figure 5 shown, the global feature extraction module includes a fifth 3×3 convolutional layer, an activation layer sub-module, a pooling layer sub-module, a fully connected layer sub-module, a flattening and transposing sub-module, a first encoder, a second encoder, a third encoder, a fourth encoder, and a fifth encoder; the structures of the first encoder, the second encoder, the third encoder, the fourth encoder, and the fifth encoder are the same, which are consistent with the encoder structure adopted in the paper "Translating Images into Maps" published on the website arXiv in October 2021, and the encoder in the paper is used to match the super-resolution network.

[0157] Input the degraded image or the reconstructed image into the global feature extraction module. The degraded image or the reconstructed image is sequentially processed by a fifth 3×3 convolutional layer, an activation layer sub-module, a pooling layer sub-module, and a fully connected layer sub-module to obtain a feature map A1. After the feature map A1 is flattened and transposed by a flattening and transposing sub-module, a feature map A2 is obtained. The feature map A2 is input into a first encoder for encoding processing to obtain a feature map A3. The feature map A3 is input into a second encoder for encoding processing to obtain a feature map A4. The feature map A4 is input into a third encoder for encoding processing to obtain a feature map A5. The feature map A5 is input into a fourth encoder for encoding processing to obtain a feature map A6. The feature map A6 is input into a fifth encoder for encoding processing to obtain a feature map A7. The attention mechanism is used to process the feature maps A3, A4, A5, A6, and A7 to correspondingly obtain the global features of the degraded image or the reconstructed image.

[0158] As Figure 6 shown, the local feature extraction module includes a first basic unit convolutional layer, a second basic unit convolutional layer, a third basic unit convolutional layer, a fourth basic unit convolutional layer, a fifth basic unit convolutional layer, a first activation layer, a second activation layer, a third activation layer, a fourth activation layer, a fifth activation layer, a first max pooling layer, a second max pooling layer, a third max pooling layer, a fourth max pooling layer, and a fifth max pooling layer; the structures of the first basic unit convolutional layer, the second basic unit convolutional layer, the third basic unit convolutional layer, the fourth basic unit convolutional layer, and the fifth basic unit convolutional layer are the same and are consistent with the basic unit convolutional layer adopted in the paper "Rethinking Depthwise Separable Convolutions: How Intra-Kernel Correlations Lead to Improved MobileNets" published in the 2020 CVPR conference.

[0159] Input the degraded image or the reconstructed image into the local feature extraction module. The degraded image or the reconstructed image is sequentially processed by the first basic unit convolutional layer, the first activation layer, and the first max pooling layer to obtain the feature map B1. The feature map B1 is sequentially processed by the second basic unit convolutional layer, the second activation layer, and the second max pooling layer to obtain the feature map B2. The feature map B2 is sequentially processed by the third basic unit convolutional layer, the third activation layer, and the third max pooling layer to obtain the feature map B3. The feature map B3 is sequentially processed by the fourth basic unit convolutional layer, the fourth activation layer, and the fourth max pooling layer to obtain the feature map B4. The feature map B4 is sequentially processed by the fifth basic unit convolutional layer, the fifth activation layer, and the fifth max pooling layer to obtain the feature map B5. The attention mechanism is used to process the feature maps B1, B2, B3, B4, and B5 to correspondingly obtain the local features of the degraded image or the reconstructed image.

[0160] As Figure 7 shown, the first basic module includes the sixth 3×3 unconstrained blueprint convolutional layer, the seventh 3×3 unconstrained blueprint convolutional layer, the eighth 3×3 unconstrained blueprint convolutional layer, the second 1×1 convolutional layer, the third 1×1 convolutional layer, the fourth 1×1 convolutional layer, the fifth 1×1 convolutional layer, the first max pooling sub-module, and the first upsampling layer. The feature map C input to the first basic module is respectively processed by the sixth 3×3 unconstrained blueprint convolutional layer and the second 1×1 convolutional layer to correspondingly obtain the feature maps C1 and C2. After the feature map C1 is respectively processed by the seventh 3×3 unconstrained blueprint convolutional layer and the first upsampling layer, the feature maps C3 and C4 are correspondingly obtained. After the feature map C3 is processed by the first max pooling sub-module, the feature map C5 is obtained. After the feature map C5 is respectively processed by the eighth 3×3 unconstrained blueprint convolutional layer and the third 1×1 convolutional layer, the feature maps C6 and C7 are correspondingly obtained. After the feature map C6 is processed by the fourth 1×1 convolutional layer, the feature map C8 is obtained. The feature maps C, C2, C4, and C8 are subjected to a splicing operation to obtain the feature map C9. After the feature map C9 is processed by the fifth 1×1 convolutional layer, the feature map C10 input to the second basic module is obtained.

[0161] As Figure 8As shown in the figure, the second basic module includes a sixth 1×1 convolutional layer, a seventh 1×1 convolutional layer, an eighth 1×1 convolutional layer, a ninth 1×1 convolutional layer, a ninth 3×3 unconstrained blueprint convolutional layer, a first 5×5 unconstrained blueprint convolutional layer, a first 7×7 unconstrained blueprint convolutional layer, a second max-pooling sub-module, a second upsampling layer, and a first Sigmoid activation layer. After the feature map C10 is processed by the sixth 1×1 convolutional layer, the feature map D1 is obtained. After the feature map D1 is processed by the seventh 1×1 convolutional layer and the second max-pooling sub-module respectively, the feature maps D2 and D3 are obtained correspondingly. After the feature map D3 is processed by the ninth 3×3 unconstrained blueprint convolutional layer, the first 5×5 unconstrained blueprint convolutional layer, and the first 7×7 unconstrained blueprint convolutional layer respectively, the feature maps D4, D5, and D6 are obtained correspondingly. The feature maps D4, D5, and D6 are input into the eighth 1×1 convolutional layer for processing to obtain the feature map D7. The feature map D7 is processed by the second upsampling layer to obtain the feature map D8. The feature map D2 and the feature map D8 are subjected to a feature addition operation to obtain the feature map D9. After the feature map D9 is processed by the ninth 1×1 convolutional layer and the first Sigmoid activation layer in sequence, the feature map D10 is obtained. The feature map D10 and the feature map C10 are subjected to a channel attention operation ( Figure 8 denoted by T in

[0162] As Figure 9 shown, based on the EuroSAT dataset, Origin is the original image; C-DM is the classical degradation model; RS-DM: the remote sensing degradation model (the degradation model of the present invention). For severely degraded remote sensing images (such as satellite images with strong atmospheric interference), the degradation model of the present invention effectively simulates the influence of atmospheric scattering through the frequency domain filtering module and performs image reconstruction using the joint spatial-frequency domain super-resolution network. Experimental results show that the reconstructed images have significant improvements in detail restoration and texture preservation compared with the prior art.

[0163] As Figures 10 - 11 shown, based on the WHU-RS19 dataset and the PatternNet dataset,

[0164] For remote sensing images with complex textures (such as remote sensing images of urban building complexes), the spatial domain branch of the present invention can effectively capture the detail features of different scales in the image, and the frequency domain branch enhances the ability to recover high-frequency information. Experimental results show that the present invention performs excellently in the reconstruction of building edges and road structures, with clear details and accurate structures.

[0165] Experiments show that the method proposed by the present invention can significantly improve the image reconstruction effect in various degradation scenarios, especially having obvious advantages in high-frequency detail and texture restoration.

[0166] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps recited in the disclosure of the present invention can be executed in parallel, sequentially, or in different orders, as long as the desired results of the technical solution disclosed in the present invention can be achieved, and no limitations are imposed herein.

[0167] The above specific embodiments do not constitute a limitation on the protection scope of the present invention. Those skilled in the art should understand that various modifications, combinations, sub - combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A remote sensing image super-resolution reconstruction method based on degradation mechanism, characterized by: The specific steps include: S1: construct a degradation model, input the high-resolution remote sensing image into the degradation model, and obtain a degraded image corresponding to the high-resolution remote sensing image; In step S1, the degradation model includes a spatial domain filtering module, a frequency domain filtering module and a degradation factor random control module, the high-resolution remote sensing image is input into the spatial domain filtering module for processing to obtain a first spatial domain image; the high-resolution remote sensing image is Fourier transformed and then input into the frequency domain filtering module, the degradation factor random control module controls the frequency domain filtering module to randomly introduce degradation factors into the high-resolution remote sensing image after Fourier transformation, and the high-resolution remote sensing image after the degradation factor is introduced is inversely Fourier transformed to obtain a second spatial domain image; the first spatial domain image and the second spatial domain image are synthesized, and then down-sampling operations and noise introduction operations are performed to obtain a degraded image; S2: construct a super-resolution network, input the degraded image into the super-resolution network, and obtain a reconstructed image; The super-resolution network includes a spatial domain branch and a frequency domain branch. The spatial domain branch extracts local multi-scale feature information of the degraded image through multiple levels of convolution operations to obtain spatial domain features. The frequency domain branch analyzes and processes the global frequency features of the degraded image to obtain frequency domain features. The spatial domain features and the frequency domain features are fused to obtain a reconstructed image; S3: Use the joint loss function to train the super-resolution network to obtain a trained super-resolution network; The joint loss function includes a pixel-level loss function, a multi-scale feature loss function and a global feature loss function; S4: Input the degraded image to be reconstructed into the trained super-resolution network to obtain a reconstructed super-resolution image.

2. The remote sensing image super-resolution reconstruction method based on degradation mechanism according to claim 1, characterized in that: The frequency domain filtering module includes an atmospheric disturbance factor, an optical system disturbance factor, a first platform component disturbance factor, a second platform component disturbance factor, and a third platform component disturbance factor; Atmospheric disturbance factor I is: ; ； in, is the angular spatial frequency, , is the focal length of the optical system, is the spatial frequency, is the refractive index structure constant, is the wavelength, R is the propagation path, a1 is the turbulence effect, a2 is the aerosol effect, is the modulation transfer function of the turbulence effect, is the modulation transfer function of the aerosol effect; ； in, is the scattering coefficient of atmospheric aerosol, is the atmospheric aerosol absorption coefficient, is the angular spatial cutoff frequency of the aerosol; The optical system disturbance factor II is: ; ； in, for , is the spatial frequency in the horizontal direction, is the spatial cutoff frequency, = , D is the diameter of the optical system, b1 is the diffraction effect, b2 is the defocus effect, b3 is the image shift effect, is the modulation transfer function of the diffraction effect, is the modulation transfer function of the defocus effect, is the modulation transfer function of the image shift effect; ； in, is a first-order Bessel function, , is the defocus amount, F is the focal ratio of the optical system, It is an intermediate parameter and has no physical meaning; ； ； in, is the image shift caused by the change in satellite velocity, is the image shift caused by the satellite yaw, is the image shift caused by the change of satellite attitude angle, is the image shift caused by the rotation of the Earth, Image shift caused by random vibration of satellite and optical remote sensor. is the image shift caused by the jitter of the integration time, It is an intermediate variable and has no physical meaning; The disturbance factor III of the first platform component is: ; ； Where N is the sampling frequency, d is the diffuse spot diameter, For low frequency vibration, It is high frequency vibration; ； in, is the zero-order Bessel function, is the platform amplitude, is the spatial frequency; The disturbance factor IV of the second platform component is: ； ； in, is the spatial frequency along the scanning direction, is the CCD pixel size, d1 is the spatial response, d2 is the temporal response, d3 is the CCD response, is the modulation transfer function of the spatial response, is the modulation transfer function of the time response, is the modulation transfer function of the CCD response; ； in, is the time frequency value along the scanning direction, is the 3-dB point frequency value of the detector time response; ； in, is the amount of charge transferred by the CCD, is the charge transfer efficiency, is the sampling frequency of the CCD signal; The third platform component disturbance factor V is: ； ； in, is the frequency value of the -3dB point of the preamplifier response, e1 is the preamplifier response, e2 is the postamplifier response, and e3 is the lift response. is the modulation transfer function of the preamplifier response, is the modulation transfer function of the post-amplifier response, is the modulation transfer function of the lift response; ； in, The frequency value of the -3dB point of the post-amplifier response; ； in, is the frequency value when the boost pressure is maximum, for The magnitude of the value at .

3. The remote sensing image super-resolution reconstruction method based on degradation mechanism according to claim 2, characterized in that: The formula for the random control module of the degradation factor to control the frequency domain filtering module to randomly introduce the degradation factor into the high-resolution remote sensing image after Fourier transformation is: K= · · ； Among them, U is uniform distribution, X1~X5 are coefficients that obey uniform distribution, n1 is 0, and n2 is 1.

4. The remote sensing image super-resolution reconstruction method based on degradation mechanism according to claim 1, characterized in that: In step S2, the spatial domain branch includes a first 3×3 unconstrained blueprint convolution layer and n cascaded basic units, each basic unit includes a cascaded first basic module and a second basic module, and the output of the first basic module is the input of the second basic module; The degraded image is copied four times, and three of the degraded images are subjected to RGB three-channel extraction operations, and the remaining degraded image is subjected to Y channel extraction operations, and all the extracted feature maps are subjected to feature splicing operations to obtain a spliced ​​feature map, and the spliced ​​feature map is subjected to a convolution operation by a first 3×3 unconstrained blueprint convolution layer to obtain an intermediate feature map, and the intermediate feature map is input into n cascaded basic units for processing to obtain a first feature map; The frequency domain branch includes a mask, a second 3×3 unconstrained blueprint convolution layer, a third 3×3 unconstrained blueprint convolution layer, a fourth 3×3 unconstrained blueprint convolution layer, a first 3×3 convolution layer, a second 3×3 convolution layer, a third 3×3 convolution layer, a first batch of normalization layers, a second batch of normalization layers, a third batch of normalization layers, a first SiLU activation function, a second SiLU activation function, and a third SiLU activation function; Performing Fourier transformation on the degraded image to obtain a frequency domain image, performing a dot multiplication operation on the frequency domain image and the mask to obtain frequency domain images under different masks, and processing the frequency domain images under different masks through an attention mechanism to obtain a second feature map, a third feature map, and a fourth feature map; The second feature map is input into the second 3×3 unconstrained blueprint convolution layer for processing to obtain the fifth feature map. The fifth feature map is processed by the first 3×3 convolution layer, the first batch of normalization layers and the first SiLU activation function in sequence to obtain the sixth feature map. The fifth feature map and the sixth feature map are feature concatenated to obtain the seventh feature map. The seventh feature map is inverse Fourier transformed to obtain the eighth feature map. The third feature map is input into the third 3×3 unconstrained blueprint convolution layer for processing to obtain the ninth feature map. The ninth feature map is processed by the second 3×3 convolution layer, the second batch of normalization layers and the second SiLU activation function in sequence. After processing, the tenth feature map is obtained, the ninth feature map and the tenth feature map are feature-concatenated to obtain the eleventh feature map, the eleventh feature map is inverse Fourier transformed to obtain the twelfth feature map; the fourth feature map is input into the fourth 3×3 unconstrained blueprint convolution layer for processing to obtain the thirteenth feature map, the thirteenth feature map is sequentially processed by the third 3×3 convolution layer, the third batch normalization layer and the third SiLU activation function to obtain the fourteenth feature map, the thirteenth feature map and the fourteenth feature map are feature-concatenated to obtain the fifteenth feature map, the fifteenth feature map is inverse Fourier transformed to obtain the sixteenth feature map.

5. The remote sensing image super-resolution reconstruction method based on degradation mechanism according to claim 4, characterized in that: The output features of the first basic module of each basic unit, the first feature map, the eighth feature map, the twelfth feature map and the sixteenth feature map are feature-concatenated to obtain a seventeenth feature map. The seventeenth feature map is sequentially convolved by the first 1×1 convolution layer and the fifth 3×3 unconstrained blueprint convolution layer to obtain an eighteenth feature map after feature supplementation. The eighteenth feature map and the intermediate feature map are added to obtain a nineteenth feature map. The nineteenth feature map is sequentially processed by the fourth 3×3 convolution layer and upsampling to obtain a reconstructed image.

6. The remote sensing image super-resolution reconstruction method based on degradation mechanism according to claim 1, characterized in that: In step S3, the output of the joint loss function is the cumulative sum of the pixel-level loss function, the multi-scale feature loss function and the global feature loss function; The pixel-level loss function is an L1 loss function, and the input of the pixel-level loss function is a degraded image and a reconstructed image; The multi-scale feature loss function is a Huber loss function with a parameter of 1, and the input of the multi-scale feature loss function is the reconstructed stitched image and the high-resolution stitched image; The global feature loss function is a Huber loss function with a parameter of 1.25, and the input of the global feature loss function is the reconstructed stitched image and the high-resolution stitched image.

7. The remote sensing image super-resolution reconstruction method based on degradation mechanism according to claim 6, characterized in that: The reconstructed stitched image is obtained by: inputting the reconstructed image into the global feature extraction module and the local feature extraction module respectively for feature extraction, and stitching the extracted global features and all local features to obtain the reconstructed stitched image; The high-resolution stitched image is obtained by inputting the high-resolution remote sensing image into the global feature extraction module and the local feature extraction module for feature extraction, and stitching the extracted global features and all local features to obtain a high-resolution stitched image.

8. The remote sensing image super-resolution reconstruction method based on degradation mechanism according to claim 7, characterized in that: The global feature extraction module includes a fifth 3×3 convolutional layer, an activation layer submodule, a pooling layer submodule, a fully connected layer submodule, a flattening transposition submodule, a first encoder, a second encoder, a third encoder, a fourth encoder, and a fifth encoder; The degraded image or the reconstructed image is input into the global feature extraction module, and the degraded image or the reconstructed image is processed by the fifth 3×3 convolutional layer, the activation layer submodule, the pooling layer submodule and the fully connected layer submodule in sequence to obtain a feature map A1. The feature map A1 is flattened and transposed by the flattening and transposing submodule to obtain a feature map A2. The feature map A2 is input into the first encoder for encoding to obtain a feature map A3. The feature map A3 is input into the second encoder for encoding to obtain a feature map A4. The feature map A4 is input into the third encoder for encoding to obtain a feature map A5. The feature map A5 is input into the fourth encoder for encoding to obtain a feature map A6. The feature map A6 is input into the fifth encoder for encoding to obtain a feature map A7. The feature maps A3, A4, A5, A6 and A7 are processed by using the attention mechanism to obtain the global features of the degraded image or the reconstructed image.

9. The remote sensing image super-resolution reconstruction method based on degradation mechanism according to claim 7, characterized in that: The local feature extraction module includes a first basic unit convolution layer, a second basic unit convolution layer, a third basic unit convolution layer, a fourth basic unit convolution layer, a fifth basic unit convolution layer, a first activation layer, a second activation layer, a third activation layer, a fourth activation layer, a fifth activation layer, a first maximum pooling layer, a second maximum pooling layer, a third maximum pooling layer, a fourth maximum pooling layer and a fifth maximum pooling layer; The degraded image or the reconstructed image is input into the local feature extraction module, and the degraded image or the reconstructed image is processed by the first basic unit convolution layer, the first activation layer and the first maximum pooling layer in sequence to obtain a feature map B1, and the feature map B1 is processed by the second basic unit convolution layer, the second activation layer and the second maximum pooling layer in sequence to obtain a feature map B2, and the feature map B2 is processed by the third basic unit convolution layer, the third activation layer and the third maximum pooling layer in sequence to obtain a feature map B3, and the feature map B3 is processed by the fourth basic unit convolution layer, the fourth activation layer and the fourth maximum pooling layer in sequence to obtain a feature map B4, and the feature map B4 is processed by the fifth basic unit convolution layer, the fifth activation layer and the fifth maximum pooling layer in sequence to obtain a feature map B5, and the feature maps B1, B2, B3, B4 and B5 are processed by the attention mechanism to obtain the local features of the degraded image or the reconstructed image.

Images