InSAR atmospheric delay correction method, device, equipment and medium

Abstract

Embodiments of the present disclosure disclose an InSAR atmospheric delay correction method, device, equipment and medium. A specific implementation of the method comprises: extracting atmospheric delay parameter data, performing data dimension processing on target elevation data to generate elevation processing data; performing interference processing on a target SAR image to generate an interferogram; for each wavelet transform mode, performing high-frequency component feature extraction and low-frequency component feature extraction on the interferogram; generating auxiliary matrix data; performing feature information fusion on at least one high-frequency component feature information and at least one low-frequency component feature information; and using an atmospheric delay correction neural network model to generate an interferogram after atmospheric delay correction. Through the predetermined data preprocessing mode and the atmospheric delay correction neural network model, the target interferogram can be quickly and accurately corrected in the atmosphere to generate a high-quality interferogram after atmospheric delay correction.

Description

Technical Field

[0001] Embodiments of this disclosure relate to the fields of computer and remote sensing technology, and more specifically to InSAR atmospheric delay correction methods, apparatus, devices, and media. Background Technology

[0002] Currently, spaceborne interferometric synthetic aperture radar (InSAR) is an advanced space-based geodesy technique. By comparing the phase information between multiple coherent radar images, it can measure minute deformation changes on the Earth's surface. This technology has been widely applied in areas such as surface deformation monitoring, plate tectonics, landslides, and glacier movement. However, the atmospheric delay effect severely restricts the further development and application of InSAR technology. Atmospheric delay refers to the difference in the path of electromagnetic waves propagating through the Earth's atmosphere due to spatiotemporal changes in atmospheric conditions (parameters such as air pressure, water vapor content, free electron density, and temperature). This difference is ultimately reflected in the phase information of the interferogram, severely interfering with the accuracy of InSAR technology in monitoring surface deformation, and in extreme cases, even causing InSAR deformation monitoring to completely fail. The common approach to addressing atmospheric delay is to establish a functional relationship between the tropospheric delay phase and the surface elevation to eliminate the influence of the tropospheric delay phase and thus solve the atmospheric delay problem.

[0003] However, when using the above methods to solve the atmospheric delay problem, the following technical problems often arise:

[0004] Solving atmospheric delay problems through functional relationships has certain limitations. It can often only correct the delay phase that is strongly correlated with the assumptions in the model, and its effect on correcting the turbulent part of atmospheric delay, which has strong spatiotemporal variability, is limited, resulting in the inability to achieve accurate atmospheric delay processing.

[0005] The information disclosed in this background section is only intended to enhance the understanding of the background of the inventive concept, and therefore may contain information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] The summary portion of this disclosure is intended to provide a brief overview of the concepts, which will be described in detail in the detailed description portion. This summary portion is not intended to identify key or essential features of the claimed technical solutions, nor is it intended to limit the scope of the claimed technical solutions.

[0007] Some embodiments of this disclosure provide InSAR atmospheric delay correction methods, apparatuses, devices, and media to address the technical problems mentioned in the background section above.

[0008] In a first aspect, some embodiments of this disclosure provide an InSAR atmospheric delay correction method, comprising: acquiring a target image, target meteorological data, and target elevation data of a target area captured by a target synthetic aperture radar; extracting atmospheric delay parameter data from the aforementioned target meteorological data, and performing data dimension processing on the aforementioned target elevation data to generate elevation-processed data; performing interferometric processing on the aforementioned target image to generate an interferogram; and, for each of at least one wavelet transform method, using the aforementioned wavelet transform method to extract high-frequency component features and low-frequency component features from the aforementioned interferogram to generate high-frequency components. Feature information and low-frequency component feature information; perform data interpolation and fusion on the above-mentioned elevation processing data and the above-mentioned atmospheric delay parameter data to generate auxiliary matrix data with the same size as the matrix corresponding to the above-mentioned interferogram; perform feature information fusion on the obtained at least one high-frequency component feature information to generate first fused feature information, and perform feature information fusion on the obtained at least one low-frequency component feature information to generate second fused feature information; input the above-mentioned first fused feature information, the above-mentioned second fused feature information and the above-mentioned auxiliary matrix data into a pre-trained atmospheric delay correction neural network model to generate an atmospheric delay correction image for the above-mentioned target image.

[0009] Secondly, some embodiments of this disclosure provide an InSAR atmospheric delay correction device, comprising: an acquisition unit configured to acquire a target image, target meteorological data, and target elevation data captured by a target synthetic aperture radar targeting a target area; a processing unit configured to extract atmospheric delay parameter data from the aforementioned target meteorological data and perform data dimension processing on the aforementioned target elevation data to generate elevation-processed data; an interferometric processing unit configured to perform interferometric processing on the aforementioned target image to generate an interferogram; and an extraction unit configured to extract high-frequency component features and low-frequency component features from the aforementioned interferogram using at least one wavelet transform method for each of the at least one wavelet transform methods, to generate... The system includes high-frequency component feature information and low-frequency component feature information; a first fusion unit configured to perform data interpolation fusion on the above-mentioned elevation processing data and the above-mentioned atmospheric delay parameter data to generate auxiliary matrix data with the same size as the matrix corresponding to the above-mentioned interferogram; a second fusion unit configured to perform feature information fusion on at least one obtained high-frequency component feature information to generate first fused feature information, and to perform feature information fusion on at least one obtained low-frequency component feature information to generate second fused feature information; and a generation unit configured to input the above-mentioned first fused feature information, the above-mentioned second fused feature information, and the above-mentioned auxiliary matrix data into a pre-trained atmospheric delay correction neural network model to generate an atmospheric delay correction image for the above-mentioned target image.

[0010] Thirdly, some embodiments of this disclosure provide an electronic device, including: one or more processors; and a storage device having one or more programs stored thereon, such that when the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any implementation of the first aspect.

[0011] Fourthly, some embodiments of this disclosure provide a computer-readable medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the method as described in any implementation of the first aspect.

[0012] The various embodiments of this disclosure have the following beneficial effects: The InSAR atmospheric delay correction method of some embodiments of this disclosure can accurately perform atmospheric correction on target images to generate high-quality synthetic aperture radar (SAR) images. Specifically, the reason for insufficient accuracy in related atmospheric corrections is that solving atmospheric delay problems through functional relationships has certain limitations. It often only corrects delay phases that are strongly correlated with assumptions in the model, and has limited effect on correcting the turbulent part of atmospheric delay, which has strong spatiotemporal variability, resulting in inaccurate atmospheric delay processing. Based on this, the InSAR atmospheric delay correction method of some embodiments of this disclosure first acquires target images, target meteorological data, and target elevation data for the target area captured by the target SAR. Here, acquiring the SAR image within the target area facilitates the subsequent acquisition of interferograms within the target area. Acquiring the target meteorological data provides meteorological information within the target area, providing meteorological-related auxiliary data during the subsequent atmospheric delay correction process. Acquiring the target elevation data provides topographic information within the target area, providing meteorological-related auxiliary data during the subsequent atmospheric delay correction process. Then, atmospheric delay parameter data is extracted from the aforementioned target meteorological data, and the target elevation data is processed to generate elevation-processed data. Here, extracting atmospheric delay parameter data from the target meteorological data allows for the extraction of key meteorological parameters. Processing the target elevation data to adjust the dimensions enables subsequent matching of phase information corresponding to the interferogram. Next, the aforementioned target image is subjected to interferometry to generate an interferogram for subsequent wavelet transform processing. For each of at least one wavelet transform method, high-frequency and low-frequency component features are extracted from the interferogram using these methods to generate high-frequency and low-frequency component feature information. Here, given the significant frequency differences in the interferogram, extracting high-frequency and low-frequency component feature information allows the subsequent atmospheric delay correction neural network model to accurately generate atmospheric delay-corrected images based on the feature information at different frequencies. The aforementioned elevation processing data and atmospheric delay parameter data are interpolated and fused to generate auxiliary matrix data of the same size as the matrix corresponding to the interferogram, which serves as auxiliary feature information for subsequent accurate generation of the atmospheric delay-corrected image. Furthermore, at least one high-frequency component feature information is fused to generate first fused feature information, and at least one low-frequency component feature information is fused to generate second fused feature information, thereby obtaining more accurately composite feature content under both high-frequency and low-frequency features.Finally, the aforementioned first fused feature information, second fused feature information, and auxiliary matrix data are input into a pre-trained atmospheric delay correction neural network model to accurately generate an atmospheric delay-corrected image for the target image. In summary, by preprocessing the target image, target meteorological data, and target elevation data, and extracting composite high-frequency and low-frequency feature information, and based on the atmospheric delay correction neural network model, an atmospheric delay-corrected image can be accurately generated. Attached Figure Description

[0013] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and elements are not necessarily drawn to scale.

[0014] Figure 1 This is a flowchart of some embodiments of the InSAR atmospheric delay correction method according to this disclosure;

[0015] Figure 2 This is a schematic diagram of the structure of some embodiments of the InSAR atmospheric delay correction device according to the present disclosure;

[0016] Figure 3 This is a schematic diagram of the structure of an electronic device suitable for implementing some embodiments of the present disclosure;

[0017] Figure 4 This is a schematic diagram illustrating model training and application of an atmospheric delay correction neural network model according to this disclosure;

[0018] Figure 5 The present disclosure illustrates a schematic diagram of the generation of the second fusion feature information.

[0019] Figure 6 The present disclosure illustrates a schematic diagram of the generation of the corrected interferogram. Detailed Implementation

[0020] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0021] It should also be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings. Unless otherwise specified, the embodiments and features described in this disclosure can be combined with each other.

[0022] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are used only to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependencies.

[0023] It should be noted that the terms "a" and "a plurality of" used in this disclosure are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0024] The names of messages or information exchanged between multiple devices in the embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

[0025] This disclosure will now be described in detail with reference to the accompanying drawings and embodiments.

[0026] refer to Figure 1 The diagram illustrates a flow 100 of some embodiments of the InSAR atmospheric delay correction method according to this disclosure. The InSAR atmospheric delay correction method includes the following steps:

[0027] Step 101: Acquire target images, meteorological data, and elevation data of the target area taken by the target synthetic aperture radar.

[0028] In some embodiments, in response to receiving session communication request information for a first user terminal and a second user terminal, the executing entity of the aforementioned InSAR atmospheric delay correction method (e.g., an electronic device) can acquire target images, target meteorological data, and target elevation data of the target area taken by a target synthetic aperture radar (SAR) via a wired or wireless connection. The target image can be an image of the target area to be atmospherically delayed and corrected by SAR. Atmospheric delay refers to the difference in the path of electromagnetic wave signals propagating through the Earth's atmosphere due to spatiotemporal changes in atmospheric conditions (parameters such as air pressure, water vapor content, free electron density, and temperature). This difference is ultimately reflected in the phase information of the interferogram, severely interfering with the accuracy of InSAR technology for monitoring surface deformation, and in extreme cases, even causing complete failure of InSAR deformation monitoring. The target meteorological data can be meteorological data currently available in the area. In practice, meteorological data within the target area can be acquired through Generic Atmospheric Correction Online Service (GACOS, an online service platform for atmospheric correction). The target meteorological data can be data with a time resolution of minutes and a spatial resolution of 90 meters. The target elevation data can be elevation data within the target area. The elevation data can be DEM (Digital Elevation Model) data. DEM data is a digital terrain model used to represent the elevation of the Earth's surface. DEM data is a model that digitally represents the topographic relief of the Earth's surface, simulating changes in ground elevation through a series of ordered numerical arrays.

[0029] Step 102: Extract atmospheric delay parameter data from the above-mentioned target meteorological data, and perform data dimension processing on the above-mentioned target elevation data to generate elevation-processed data.

[0030] In some embodiments, the aforementioned executing entity can extract atmospheric delay parameter data from the aforementioned target meteorological data and perform data dimension processing on the aforementioned target elevation data to generate elevation-processed data. The atmospheric delay parameter data can be at least one parameter that has a significant impact on atmospheric delay. In practice, atmospheric delay parameter data can include water vapor content, temperature, and air pressure parameters. Data dimension processing can involve performing a phase-to-elevation conversion on the elevation data to unify the dimensions of the elevation data with the dimensions of the phase data in the interferogram. That is, the dimensions corresponding to the elevation-processed data are the same as the dimensions of the phase data in the interferogram.

[0031] In some optional implementations of certain embodiments, the execution entity may extract atmospheric delay parameter data from the target meteorological data and perform data dimension processing on the target elevation data to generate elevation-processed data, including the following steps:

[0032] The first step is to perform the following generation steps for each raster data in the target elevation data:

[0033] Sub-step 1: Divide the target value by the radar wavelength to generate the first division result. The target value can be 4π. The radar wavelength refers to the wavelength of the electromagnetic waves used in the radar system. The target elevation data can be in raster format; that is, the target elevation data includes multiple raster data.

[0034] Sub-step 2 involves multiplying the satellite distance corresponding to the aforementioned grid data by the sine value corresponding to the radar wave incident angle to obtain the multiplication result. Here, the aforementioned satellite distance can be the distance between the satellite and the ground corresponding to the grid data.

[0035] Sub-step 3 involves dividing the vertical baseline length corresponding to the above grid data by the multiplication result to obtain the second division result. The vertical baseline length refers to the distance component between two antennas (or radar sensors) in the vertical line-of-sight direction (i.e., the direction perpendicular to the ground target) in an InSAR system. The vertical baseline length is a value relative to the parallel baseline (i.e., the component of the baseline in the line-of-sight slant range direction).

[0036] Sub-step 4 involves multiplying the first division result, the second division result, and the ground height corresponding to the grid data to obtain the processed grid data. The ground height can be the height of the ground protrusion corresponding to the grid data.

[0037] The second step is to identify the processed raster dataset as the elevation data described above.

[0038] Step 103: Perform interferometric processing on the target image to generate an interferogram.

[0039] In some embodiments, the aforementioned execution entity can perform interference processing on the target image using conventional interference processing methods to generate an interference map.

[0040] Step 104: For each wavelet transform method in at least one wavelet transform method, the high-frequency component feature extraction and low-frequency component feature extraction are performed on the interferogram using the wavelet transform method to generate high-frequency component feature information and low-frequency component feature information.

[0041] In some embodiments, the execution entity can extract high-frequency and low-frequency features from the interferogram using each of the at least one wavelet transform methods to generate high-frequency and low-frequency feature information. The high-frequency feature information can be the feature content corresponding to the high-frequency signal in the interferogram. The low-frequency feature information can be the feature content corresponding to the low-frequency signal in the interferogram. The at least one wavelet transform method can include, but is not limited to, at least one of the following: Daubechies wavelet function, Coiflet (Coiflet Wavelet) wavelet function, and Haar (Haar Wavelet Basis Function) wavelet basis function. Among these, using the Haar wavelet basis function not only provides good time resolution, but also shows greater sensitivity to processing abrupt signals in atmospheric delays, which often contain abrupt changes. The Daubechies wavelet function has higher smoothness, enabling it to process smooth signals, corresponding to the stationary portion of atmospheric delay signals. Furthermore, the Daubechies wavelet function has stronger analytical capabilities for signals rich in detail, preserving rich detail information in atmospheric delay data. The Coiflet wavelet function, due to its superior smoothness, is suitable for analyzing complex signals. Considering regions with rapidly changing meteorological conditions and exceptionally complex atmospheric delay patterns, the Coiflet wavelet basis function also possesses unique advantages in capturing complex signal patterns. High-frequency component features provide rich details of atmospheric delay within the interferogram. Low-frequency component features provide large-scale atmospheric delay profile background information.

[0042] Here, considering the significant differences in the atmospheric delay signals in the frequency domain of the acquired interferograms, preprocessing is used to separate the significant signals to improve the accuracy of subsequent atmospheric correction.

[0043] Step 105: Perform data interpolation and fusion on the above-mentioned elevation processing data and the above-mentioned atmospheric delay parameter data to generate auxiliary matrix data with the same size as the matrix corresponding to the above-mentioned interferogram.

[0044] In some embodiments, the execution entity may perform data interpolation and fusion on the elevation processing data and the atmospheric delay parameter data to generate auxiliary matrix data with the same size as the matrix corresponding to the interferogram.

[0045] Step 106: The obtained at least one high-frequency component feature information is fused to generate first fused feature information, and the obtained at least one low-frequency component feature information is fused to generate second fused feature information.

[0046] In some embodiments, the execution entity may fuse at least one high-frequency component feature information to generate first fused feature information, and fuse at least one low-frequency component feature information to generate second fused feature information.

[0047] As an example, the aforementioned executing entity can perform feature information averaging on at least one high-frequency component feature information to generate first fused feature information, and perform feature information averaging on at least one low-frequency component feature information to generate second fused feature information.

[0048] Step 107: Input the first fusion feature information, the second fusion feature information, and the auxiliary matrix data into the pre-trained atmospheric delay correction neural network model to generate an atmospheric delay correction image for the target image.

[0049] In some embodiments, the executing entity may input the first fused feature information, the second fused feature information, and the auxiliary matrix data into a pre-trained atmospheric delay correction neural network model to generate an atmospheric delay-corrected image for the target image. The atmospheric delay-corrected image may be a synthetic aperture radar image after atmospheric delay correction. The atmospheric delay correction neural network model may be a neural network model performing atmospheric delay correction operations. For example, the atmospheric delay correction neural network model may be a multi-layered cascaded convolutional layer.

[0050] In some optional implementations of certain embodiments, the atmospheric delay correction neural network model includes: a multi-branch feature encoding extractor, a feature fusion network, and a multi-scale regression network. The multi-branch feature encoding extractor can be a feature encoding extractor with multiple branches. The feature encoding extractor can be an encoder that extracts semantic content of features. The multi-branch feature encoding extractor can include: a first feature encoding extractor corresponding to the first fused feature information, a second feature encoding extractor corresponding to the second fused feature information, and a third feature encoding extractor corresponding to the auxiliary matrix data. The number of convolutional layers corresponding to the first, second, and third feature encoding extractors can be different, and the number of layers can be adjusted adaptively. The feature fusion network can be a network layer that fuses feature information from various input feature information. The multi-branch feature encoding extractor extracts the high-frequency detail component channel, the low-frequency background component channel, and the meteorological and topographic information auxiliary component channel from the input data. Considering the multi-scale information present in the atmospheric delay features, as well as the large-scale transformation information and the local change information of small areas, a corresponding multi-scale regression network is set to perform corresponding multi-scale regression analysis on the features to accurately capture the changes in the corresponding patterns. Multi-scale regression networks contain different branches with convolutional kernel sizes of 1, 3, and 5. By using kernels of different sizes to process information at different scales, they acquire richer feature information at various scales. Finally, the features acquired from different branches are fused, normalized, and activated to obtain the model's final output features.

[0051] Optionally, the execution entity may input the first fusion feature information, the second fusion feature information, and the auxiliary matrix data into a pre-trained atmospheric delay correction neural network model to generate an atmospheric delay correction image for the target image, including the following steps:

[0052] The first step involves inputting the aforementioned first fused feature information, the aforementioned second fused feature information, and the aforementioned auxiliary matrix data into the aforementioned multi-branch feature encoding extractor to generate a multi-channel feature information set. The multi-channel feature information set may include: feature encoding information corresponding to the first fused feature information, feature encoding information corresponding to the second fused feature information, and feature encoding information corresponding to the auxiliary matrix data.

[0053] The second step is to input the multi-channel feature information set into the feature fusion network to generate fused feature information.

[0054] The third step is to input the aforementioned fused feature information into the aforementioned multi-scale regression network to generate the aforementioned atmospheric delay-corrected image.

[0055] In some optional implementations of certain embodiments, the multi-branch feature encoding extractor described above includes: a low-frequency feature extraction module, a high-frequency feature extraction module, and an auxiliary data feature extraction module. The low-frequency feature extraction module can be a network layer that extracts low-frequency semantic content from low-frequency signals. The high-frequency feature extraction module can be a network layer that extracts high-frequency semantic content from high-frequency signals. The auxiliary data feature extraction module can be a network layer that extracts specific semantic content from auxiliary data. The low-frequency feature extraction module is used to extract large-scale contour information in the interferometric phase image, requiring branches with large-scale feature capture capabilities. Therefore, the basic module of this low-frequency feature extraction module consists of a 7x7 convolutional kernel, a normalization layer, an activation function, and a pooling layer in sequence. Since low-frequency feature information is mostly large-scale, contour-level information, the features are relatively simple and do not require a deep network structure to extract high-level abstract features. Therefore, considering the balance between computational efficiency and fitting accuracy, the basic module of the low-frequency feature extraction module is repeated three times to obtain the low-frequency feature extraction module. Since high-frequency signals mainly contain interferometric detail variation information, on the one hand, the spatial variation scale of the detail information is small, and a large convolutional kernel is not required. On the other hand, the detailed information contains a large amount of data, requiring a deeper network structure to extract more abstract and advanced robust features. The high-frequency feature extraction module is optimized using a 3x3 kernel, a deeper network structure, and a larger number of basic channels. The basic module of the high-frequency feature extraction module consists of a 3×3 convolutional layer, a batch normalization layer (BN), a Gaussian error linear unit (Gelu), and a channel attention layer. The high-frequency feature extraction module uses a basic channel configuration of 64, and the basic module has a repetition count (depth) of 5. The input data for the auxiliary data feature extraction module not only has high spatial resolution but also significantly increased data complexity. Therefore, the auxiliary data feature extraction module uses a combination of dilated convolution, batch normalization, Gaussian activation, and global average pooling as the basic module for the auxiliary information branch. Among these, dilated convolution can effectively expand the receptive field, maintain data resolution, and perform multi-scale feature extraction, making full and efficient use of the input data. The input data of the auxiliary data feature extraction module contains a large amount of information. Therefore, the basic number of channels corresponding to the auxiliary data feature extraction module is 64. The configuration of this basic module is repeated 5 times to build a large branch structure to obtain high-level abstract features. After each basic module is executed, a channel attention mechanism module is used to adjust the importance of different feature channels and determine the optimal weight ratio between the target image, target meteorological data and target elevation data to ensure the accuracy of subsequent atmospheric correction.

[0056] Optionally, the execution entity may input the first fused feature information, the second fused feature information, and the auxiliary matrix data into the multi-branch feature encoder extractor to generate a multi-channel feature information set, including the following steps:

[0057] The first step is to input the second fusion feature information into the low-frequency feature extraction module to generate low-frequency channel feature information.

[0058] The second step is to input the first fused feature information into the high-frequency feature extraction module to generate high-frequency channel feature information.

[0059] The third step is to input the aforementioned auxiliary matrix data into the aforementioned auxiliary data feature extraction module to generate auxiliary channel feature information.

[0060] The fourth step is to generate the multi-channel feature information set based on the low-frequency channel feature information, the high-frequency channel feature information, and the auxiliary channel feature information.

[0061] As an example, the aforementioned execution entity can directly combine the aforementioned low-frequency channel feature information, the aforementioned high-frequency channel feature information, and the aforementioned auxiliary channel feature information to generate a multi-channel feature information set.

[0062] In some optional implementations of certain embodiments, the execution entity may input the multi-channel feature information set into the feature fusion network to generate fused feature information, including the following steps:

[0063] The first step is to input the multi-channel feature information set into the average pooling layer of the feature fusion network to generate the average pooling result.

[0064] The second step is to input the above average pooling result into the attention weight calculation layer included in the above feature fusion network to generate attention feature information.

[0065] The third step is to input the attention feature information into the multilayer perceptron included in the feature fusion network to generate a feedforward output.

[0066] The fourth step is to input the above feedforward output results into the batch normalization layer included in the above feature fusion network to generate normalization results.

[0067] The fifth step involves inputting the normalization result into the activation function layer of the feature fusion network to generate the fused feature information.

[0068] In some optional implementations of certain embodiments, the execution entity may input the fused feature information into the multi-scale regression network to generate the atmospheric delay-corrected image, including the following steps:

[0069] The first step is to input the aforementioned fused feature information into the convolutional layer at the first convolutional kernel scale of the multi-scale regression network to generate the first convolution result.

[0070] The second step is to input the aforementioned fused feature information into the convolutional layer at the second convolutional kernel scale of the multi-scale regression network to generate a second convolutional result, wherein the second convolutional kernel scale is higher than the first convolutional kernel scale.

[0071] The third step is to input the aforementioned fused feature information into the convolutional layer at the third convolutional kernel scale of the aforementioned multi-scale regression network to generate the third convolution result, wherein the aforementioned third convolutional kernel scale is higher than the aforementioned second convolutional kernel scale.

[0072] The fourth step is to input the aforementioned fused feature information into the convolutional layer at the fourth convolutional kernel scale of the aforementioned multi-scale regression network to generate the fourth convolution result, wherein the aforementioned fourth convolutional kernel scale is higher than the aforementioned third convolutional kernel scale.

[0073] The fifth step is to input the second convolution result into the convolutional layer at the first convolutional kernel scale of the multi-scale regression network to generate the fifth convolution result.

[0074] The sixth step is to input the third convolution result into the convolutional layer at the first convolutional kernel scale of the multi-scale regression network to generate the sixth convolution result.

[0075] Step 7: Input the fourth convolution result into the convolutional layer at the first convolutional kernel scale of the multi-scale regression network to generate the seventh convolution result.

[0076] The eighth step is to combine the results of the first, fifth, sixth, and seventh convolutions to generate combined feature information.

[0077] The ninth step involves inputting the combined feature information described above into the batch normalization layer and the activation function layer to output the atmospheric delay-corrected image.

[0078] In some alternative implementations of certain embodiments, the atmospheric delay correction neural network model is trained through the following steps:

[0079] The first step involves acquiring raw synthetic aperture radar (SAR) image sets, meteorological data datasets, and elevation datasets under different atmospheric imaging environments. The raw SAR image set can be an image set under various imaging meteorological conditions to ensure it encompasses a diverse and rich variety of atmospheric delay signal samples. Simultaneously, the raw SAR image set can be an image set under different imaging region type factors. For example, different imaging region type factors could include mountainous areas, plains, etc. These different regions exhibit significant differences in atmospheric delay signal patterns. Sample sets created in relevant regions help enhance the model's atmospheric delay correction results under various terrain conditions. The meteorological data dataset can utilize GACOS meteorological data.

[0080] The second step involves interferometric processing of the original synthetic aperture radar (SAR) images in the aforementioned original SAR image set to generate radar interferograms, thus obtaining the radar interferogram atlas. The specific implementation details are omitted here.

[0081] Third, for each radar interferogram in the above radar interferogram set, perform the following generation steps:

[0082] Sub-step 1: For each of the at least one wavelet transform methods, the high-frequency component features and low-frequency component features of the radar interferogram are extracted using the aforementioned wavelet transform methods to generate radar high-frequency component feature information and radar low-frequency component feature extraction information. The specific implementation method will not be elaborated further.

[0083] Sub-step 2 involves fusing the obtained high-frequency radar component feature information to generate third fused feature information, and fusing the obtained low-frequency radar component feature information to generate fourth fused feature information.

[0084] Sub-step 3 involves preprocessing the corresponding meteorological and elevation data to generate auxiliary matrix data. This preprocessing may include extracting atmospheric delay parameters from the meteorological data and performing dimensionality processing on the target elevation data.

[0085] Sub-step 4: Combine the above-mentioned third fusion feature information, the above-mentioned fourth fusion feature information and the above-mentioned auxiliary matrix data to generate combined data.

[0086] Sub-step 5 involves applying at least one atmospheric correction method to the aforementioned radar interferogram to generate at least one atmospherically corrected interferogram. The atmospheric correction method can be an atmospheric delay correction method. In practice, at least one atmospheric correction method may include: methods based on external meteorological data, WRF (Weather Research and Forecasting Model) atmospheric models, methods based on phase-elevation empirical relationships, methods based on MERIS (Medium Resolution Imaging Spectrometer), and atmospheric delay correction methods based on spatiotemporal filtering. Methods based on external meteorological data and WRF atmospheric models show better correction results when meteorological changes are strong. Atmospheric delay correction methods based on phase-elevation empirical relationships, MERIS spectrometers, and spatiotemporal filtering methods show good correction results for atmospheric delay caused by altitude.

[0087] Sub-step 6 involves determining the terrain roughness (TR) and atmospheric variability (AV) information corresponding to the aforementioned radar interferogram. The terrain roughness information characterizes the average terrain roughness within the region corresponding to the radar interferogram. The atmospheric variability information characterizes the degree of atmospheric variation within the region.

[0088] As an example, the aforementioned implementing entity can use the standard deviation calculation formula to determine the terrain undulation information and atmospheric change intensity information corresponding to the aforementioned radar interferogram.

[0089] Sub-step 7: Based on the aforementioned terrain undulation information and atmospheric change intensity information, determine at least one correction weight corresponding to the aforementioned at least one atmospheric correction method. The correction weight characterizes the effectiveness of the atmospheric correction method, i.e., the degree to which it is subsequently used. The correction weight can be a value between 0 and 1. A higher value indicates greater effectiveness of the atmospheric correction method. Each atmospheric correction method has a corresponding correction weight.

[0090] As an example, the aforementioned implementing entity can match at least one correction weight corresponding to at least one atmospheric correction method based on the intervals corresponding to the terrain undulation information and the atmospheric change intensity information.

[0091] Sub-step 8: Based on at least one of the above-mentioned correction weights, perform interferogram fusion on the at least one atmospheric correction interferogram to generate a correction interferogram.

[0092] As an example, the aforementioned implementing entity can perform weighted fusion of at least one atmospheric correction interferogram based on at least one correction weight to generate a correction interferogram.

[0093] Sub-step 9: Use the above combined data as training data, use the above corrected interferogram as interferogram label, and combine the above combined data and the above corrected interferogram to generate training data.

[0094] The fourth step involves training the initial atmospheric delay correction neural network model using the obtained training dataset to generate a new atmospheric delay correction neural network model. The initial atmospheric delay correction neural network model can be an incomplete training model.

[0095] As an example, the aforementioned execution entity can use conventional model training methods (such as backpropagation-based model parameter update methods) to train the initial atmospheric delay correction neural network model based on the obtained training dataset to generate an atmospheric delay correction neural network model.

[0096] The aforementioned "steps one through four" are one of the inventive points of this disclosure, solving another technical problem: "how to accurately train the atmospheric delay correction neural network to obtain a more accurate model for atmospheric delay correction." Based on this, this disclosure preprocesses and extracts features from the basic data by performing interferometric processing on the original synthetic aperture radar image, applying various transformations to the radar interferogram, and extracting feature information and elevation data at each feature level for data preprocessing. At least one atmospheric correction method is used to generate atmospheric correction maps under each method, and the correction weights corresponding to each atmospheric correction method are determined based on the degree of terrain undulation and the intensity of atmospheric changes, in order to accurately correct the interferogram. Based on the above processing methods, training data can be accurately generated to accurately train the initial atmospheric delay correction neural network model, resulting in a more accurate model for atmospheric delay correction.

[0097] Optionally, the implementing entity may determine at least one correction weight corresponding to the at least one atmospheric correction method based on the aforementioned terrain undulation information and the aforementioned atmospheric change intensity information, which may include the following steps:

[0098] The first step is to multiply the information on the degree of atmospheric change mentioned above with the target value to obtain the first multiplication result. The target value can be a pre-set value. For example, the target value could be 4.

[0099] The second step is to multiply the above terrain undulation information with the target value to obtain the second multiplication result.

[0100] Third, in response to determining that the above-mentioned atmospheric change severity information is greater than or equal to the above-mentioned second multiplication result, the correction weight ratio corresponding to the above-mentioned at least one atmospheric correction method is determined as the first correction weight ratio. The first correction weight ratio can be a pre-set ratio value. For example, the first correction weight ratio can be "4:2:2:1:1".

[0101] The fourth step is to normalize the first correction weight ratio to generate at least one correction weight corresponding to the at least one atmospheric correction method.

[0102] Fifth, in response to determining that the second multiplication result is greater than the atmospheric change severity information and the atmospheric change severity information is greater than the topographic relief information, a correction weight ratio corresponding to the at least one atmospheric correction method is determined as a second correction weight ratio. This second correction weight ratio can be a pre-set ratio value. For example, the second correction weight ratio can be "3:2:3:1:1".

[0103] The sixth step is to normalize the second correction weight ratio to generate at least two correction weights corresponding to the at least one atmospheric correction method.

[0104] Step 7: In response to determining that the first multiplication result is greater than or equal to the terrain undulation information, determine the correction weight ratio corresponding to the at least one atmospheric correction method as the third correction weight ratio. The third correction weight ratio can be a pre-set ratio value. For example, the third correction weight ratio can be "5:2:1:1:1".

[0105] The eighth step is to normalize the third correction weight ratio to generate at least one correction weight corresponding to the above-mentioned atmospheric correction method.

[0106] In the ninth step, in response to determining that the first multiplication result is greater than the terrain undulation information and the terrain undulation information is greater than the atmospheric change severity information, the correction weight ratio corresponding to the at least one atmospheric correction method is determined as the fourth correction weight ratio. The fourth correction weight ratio can be a pre-set ratio value. For example, the fourth correction weight ratio can be "4:2:2:1:1".

[0107] The aforementioned "Steps 1-9" are one of the inventive points of this disclosure, solving another technical problem: "how to accurately fuse the atmospheric correction results corresponding to at least one atmospheric correction method to obtain accurate atmospheric correction labels." Based on this, this disclosure analyzes two indicators—topographic relief information and atmospheric change intensity information—to determine the main sources of atmospheric delay in the sample library, and then allocates weight ratios for different methods accordingly. The principle of weight allocation is that when atmospheric delay caused by drastic meteorological changes is dominant, the improved method based on external meteorological data and meteorological models is given a larger integration weight; when atmospheric delay is dominated by topographic relief, the method based on phase-elevation relationships is designed with a larger weight. Thus, at least one correction weight can be accurately generated under various conditions, and the fusion of the output results corresponding to at least one atmospheric correction method can be achieved to obtain accurate atmospheric correction labels.

[0108] like Figure 4 The diagram illustrates the training and application of an atmospheric delay correction neural network model.

[0109] like Figure 4 As shown, after preprocessing the SAR interferogram, meteorological data (corresponding to meteorological data), and DEM data (corresponding to elevation data) corresponding to the original synthetic aperture radar image, the data are sequentially input into a multi-branch feature extraction module (corresponding to a multi-branch feature encoding extractor), a channel attention mechanism fusion module (feature fusion network), and a multi-scale regression head (multi-scale regression network) to generate an atmospherically corrected predicted image (i.e., Data_pred). In the application phase, an atmospherically corrected image (i.e., Data) is directly generated. During the generation of interferogram labels, firstly, for the SAR interferogram, corrected data1, corrected data2, and corrected data3 are generated using filtering methods, phase-elevation relationship methods, and meteorological models, respectively. Then, based on the weights corresponding to at least one correction method, and the corrected data1, corrected data2, and corrected data3, interferogram labels are generated. Based on the interferogram labels and the atmospherically corrected predicted image, model training for an atmospheric delay correction neural network model can be achieved.

[0110] like Figure 5 As shown, a schematic diagram illustrating the generation of the second fused feature information is presented.

[0111] like Figure 6 As shown, a schematic diagram of the generation of the corrected interferogram is illustrated.

[0112] The above-described embodiments of this disclosure have the following beneficial effects: The atmospheric correction methods of some embodiments of this disclosure can accurately correct target images to generate high-quality synthetic aperture radar (SAR) images. Specifically, the reason for insufficient accuracy in related atmospheric corrections is that solving atmospheric delay problems through functional relationships has certain limitations. It often only corrects delay phases that are strongly correlated with assumptions in the model, and has limited effect on correcting the turbulent portion of atmospheric delay, which has strong spatiotemporal variability, resulting in inaccurate atmospheric delay processing. Based on this, the large InSAR atmospheric delay correction method of some embodiments of this disclosure first acquires target images, target meteorological data, and target elevation data for the target area captured by the target SAR. Here, acquiring the SAR image within the target area facilitates the subsequent acquisition of interferograms within the target area. Acquiring the target meteorological data provides meteorological information within the target area, providing meteorological-related auxiliary data during the subsequent atmospheric delay correction process. Acquiring the target elevation data provides topographic information within the target area, providing meteorological-related auxiliary data during the subsequent atmospheric delay correction process. Then, atmospheric delay parameter data is extracted from the aforementioned target meteorological data, and the target elevation data is processed to generate elevation-processed data. Here, extracting atmospheric delay parameter data from the target meteorological data allows for the extraction of key meteorological parameters. Processing the target elevation data to adjust the dimensions enables subsequent matching of phase information corresponding to the interferogram. Next, the aforementioned target image is subjected to interferometry to generate an interferogram for subsequent wavelet transform processing. For each of at least one wavelet transform method, high-frequency and low-frequency component features are extracted from the interferogram using these methods to generate high-frequency and low-frequency component feature information. Here, given the significant frequency differences in the interferogram, extracting high-frequency and low-frequency component feature information allows the subsequent atmospheric delay correction neural network model to accurately generate atmospheric delay-corrected images based on the feature information at different frequencies. The aforementioned elevation processing data and atmospheric delay parameter data are interpolated and fused to generate auxiliary matrix data of the same size as the matrix corresponding to the interferogram, which serves as auxiliary feature information for subsequent accurate generation of the atmospheric delay-corrected image. Furthermore, at least one high-frequency component feature information is fused to generate first fused feature information, and at least one low-frequency component feature information is fused to generate second fused feature information, thereby obtaining more accurately composite feature content under both high-frequency and low-frequency features.Finally, the aforementioned first fused feature information, second fused feature information, and auxiliary matrix data are input into a pre-trained atmospheric delay correction neural network model to accurately generate an atmospheric delay-corrected image for the target image. In summary, by preprocessing the target image, target meteorological data, and target elevation data, and extracting composite high-frequency and low-frequency feature information, and based on the atmospheric delay correction neural network model, an atmospheric delay-corrected image can be accurately generated.

[0113] Further reference Figure 2 As an implementation of the methods shown in the above figures, this disclosure provides some embodiments of an InSAR atmospheric delay correction device, which are similar to... Figure 1 Corresponding to the method embodiments shown, this InSAR atmospheric delay correction device can be specifically applied to various electronic devices.

[0114] like Figure 2 As shown, an InSAR atmospheric delay correction device 200 includes: an acquisition unit 201, a processing unit 202, an interferometric processing unit 203, an extraction unit 204, a first fusion unit 205, a second fusion unit 206, and a generation unit 207. The acquisition unit 201 is configured to acquire target images, target meteorological data, and target elevation data captured by a target synthetic aperture radar targeting a target area. The processing unit 202 is configured to extract atmospheric delay parameter data from the aforementioned target meteorological data and to perform data dimension processing on the aforementioned target elevation data to generate elevation-processed data. The interferometric processing unit 203 is configured to perform interferometric processing on the aforementioned target images to generate interferograms. The extraction unit 204 is configured to extract high-frequency component features and low-frequency component features from the interferograms using at least one wavelet transform method for each of the at least one wavelet transform methods, thereby generating high-frequency component feature information and low-frequency component feature information. Information; a first fusion unit 205 is configured to perform data interpolation fusion on the above-mentioned elevation processing data and the above-mentioned atmospheric delay parameter data to generate auxiliary matrix data with the same size as the matrix corresponding to the above-mentioned interferogram; a second fusion unit 206 is configured to perform feature information fusion on the obtained at least one high-frequency component feature information to generate first fused feature information, and perform feature information fusion on the obtained at least one low-frequency component feature information to generate second fused feature information; a generation unit 207 is configured to input the above-mentioned first fused feature information, the above-mentioned second fused feature information and the above-mentioned auxiliary matrix data into a pre-trained atmospheric delay correction neural network model to generate an atmospheric delay correction image for the above-mentioned target image.

[0115] It is understandable that the elements described in the InSAR atmospheric delay correction device 200 are related to the reference. Figure 1 The steps in the described method correspond to each other. Therefore, the operations, characteristics, and beneficial effects described above for the method also apply to the InSAR atmospheric delay correction device 200 and the units contained therein, and will not be repeated here.

[0116] The following is for reference. Figure 3 It shows a schematic diagram of the structure of an electronic device (e.g., an electronic device) 300 suitable for implementing some embodiments of the present disclosure. Figure 3 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of this disclosure.

[0117] like Figure 3 As shown, the electronic device 300 may include a processing unit (e.g., a central processing unit, a graphics processing unit, etc.) 301, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 302 or a program loaded from a storage device 308 into a random access memory (RAM) 303. The RAM 303 also stores various programs and data required for the operation of the electronic device 300. The processing unit 301, ROM 302, and RAM 303 are interconnected via a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.

[0118] Typically, the following devices can be connected to I / O interface 305: input devices 306 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 307 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 308 including, for example, magnetic tapes, hard disks, etc.; and communication devices 309. Communication device 309 allows electronic device 300 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 3 An electronic device 300 with various devices is shown; however, it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed alternatively. Figure 3 Each box shown can represent a device or multiple devices as needed.

[0119] In particular, according to some embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, some embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication device 309, or installed from storage device 308, or installed from ROM 302. When the computer program is executed by processing device 301, it performs the functions defined in the methods of some embodiments of this disclosure.

[0120] It should be noted that, in some embodiments of this disclosure, the computer-readable medium described above may be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In some embodiments of this disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In some embodiments of this disclosure, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0121] In some implementations, clients and servers can communicate using any currently known or future-developed network protocol such as HTTP (Hypertext Transfer Protocol) and can interconnect with digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (“LANs”), wide area networks (“WANs”), the Internet (e.g., the Internet of Things), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future-developed networks.

[0122] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device. The aforementioned computer-readable medium carries one or more programs that, when executed by the electronic device, cause the electronic device to: acquire target images, target meteorological data, and target elevation data of a target area captured by a target synthetic aperture radar; extract atmospheric delay parameter data from the aforementioned target meteorological data, and perform data dimension processing on the aforementioned target elevation data to generate elevation-processed data; perform interferometric processing on the aforementioned target images to generate interferograms; and, for each of at least one wavelet transform method, use the aforementioned wavelet transform method to extract high-frequency component features and low-frequency component features from the aforementioned interferograms. The process involves generating high-frequency component feature information and low-frequency component feature information; performing data interpolation and fusion on the aforementioned elevation processing data and the aforementioned atmospheric delay parameter data to generate auxiliary matrix data with the same size as the matrix corresponding to the aforementioned interferogram; fusing the obtained at least one high-frequency component feature information to generate first fused feature information, and fusing the obtained at least one low-frequency component feature information to generate second fused feature information; and inputting the aforementioned first fused feature information, the aforementioned second fused feature information, and the aforementioned auxiliary matrix data into a pre-trained atmospheric delay correction neural network model to generate an atmospheric delay correction image for the aforementioned target image.

[0123] Computer program code for performing operations of some embodiments of this disclosure can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0124] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0125] The units described in some embodiments of this disclosure can be implemented in software or hardware. The described units can also be housed in a processor; for example, a processor may be described as including an acquisition unit, a processing unit, an interferometric processing unit, an extraction unit, a first fusion unit, a second fusion unit, and a generation unit. The names of these units do not necessarily limit the specific unit itself; for example, the acquisition unit may also be described as "a unit that acquires target images, target meteorological data, and target elevation data of a target area captured by a target synthetic aperture radar."

[0126] The functions described above in this document can be performed at least in part by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), system-on-a-chip (SoCs), complex programmable logic devices (CPLDs), and so on.

[0127] The above description is merely a selection of preferred embodiments of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in the embodiments of this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions disclosed in the embodiments of this disclosure.

Claims

1. An InSAR atmospheric delay correction method, comprising: Acquire target images, meteorological data, and elevation data of the target area captured by the target synthetic aperture radar; Atmospheric delay parameter data is extracted from the target meteorological data, and the target elevation data is processed in terms of data dimensions to generate elevation-processed data; The target image is subjected to interferometric processing to generate an interferogram; For each wavelet transform method in at least one wavelet transform method, the high-frequency component feature extraction and low-frequency component feature extraction are performed on the interferogram using the wavelet transform method to generate high-frequency component feature information and low-frequency component feature information; The elevation processing data and the atmospheric delay parameter data are interpolated and fused to generate auxiliary matrix data with the same size as the matrix corresponding to the interferogram; The feature information of at least one high-frequency component obtained is fused to generate first fused feature information, and the feature information of at least one low-frequency component obtained is fused to generate second fused feature information. The first fused feature information, the second fused feature information, and the auxiliary matrix data are input into a pre-trained atmospheric delay correction neural network model to generate an atmospheric delay correction image for the target image.

2. The method according to claim 1, wherein, The atmospheric delay correction neural network model includes: a multi-branch feature encoding extractor, a feature fusion network, and a multi-scale regression network; and The step of inputting the first fused feature information, the second fused feature information, and the auxiliary matrix data into a pre-trained atmospheric delay correction neural network model to generate an atmospheric delay correction image for the target image includes: The first fused feature information, the second fused feature information, and the auxiliary matrix data are input into the multi-branch feature encoder extractor to generate a multi-channel feature information set; The multi-channel feature information set is input into the feature fusion network to generate fused feature information; The fused feature information is input into the multi-scale regression network to generate the atmospheric delay-corrected interferometric image.

3. The method according to claim 2, wherein, The multi-branch feature encoding extractor includes: a low-frequency feature extraction module, a high-frequency feature extraction module, and an auxiliary data feature extraction module; and The step of inputting the first fused feature information, the second fused feature information, and the auxiliary matrix data into the multi-branch feature encoder to generate a multi-channel feature information set includes: The second fused feature information is input into the low-frequency feature extraction module to generate low-frequency channel feature information; The first fused feature information is input into the high-frequency feature extraction module to generate high-frequency channel feature information; The auxiliary matrix data is input into the auxiliary data feature extraction module to generate auxiliary channel feature information; The multi-channel feature information set is generated based on the low-frequency channel feature information, the high-frequency channel feature information, and the auxiliary channel feature information.

4. The method according to claim 2, wherein, The step of inputting the multi-channel feature information set into the feature fusion network to generate fused feature information includes: The multi-channel feature information set is input into the average pooling layer included in the feature fusion network to generate the average pooling result; The average pooling result is input into the attention weight calculation layer of the feature fusion network to generate attention feature information; The attention feature information is input into the multilayer perceptron included in the feature fusion network to generate a feedforward output result; The feedforward output is input into the batch normalization layer of the feature fusion network to generate a normalized result. The normalization result is input into the activation function layer of the feature fusion network to generate the fused feature information; and The step of inputting the fused feature information into the multi-scale regression network to generate the atmospheric delay-corrected image includes: The fused feature information is input into the convolutional layer at the first convolutional kernel scale of the multi-scale regression network to generate the first convolution result. The fused feature information is input into the convolutional layer at the second convolutional kernel scale of the multi-scale regression network to generate a second convolutional result, wherein the second convolutional kernel scale is higher than the first convolutional kernel scale; The fused feature information is input into the convolutional layer at the third convolutional kernel scale of the multi-scale regression network to generate a third convolution result, wherein the third convolutional kernel scale is higher than the second convolutional kernel scale; The fused feature information is input into the convolutional layer at the fourth convolutional kernel scale of the multi-scale regression network to generate the fourth convolution result, wherein the fourth convolutional kernel scale is higher than the third convolutional kernel scale; The second convolution result is input into the convolutional layer at the first convolutional kernel scale of the multi-scale regression network to generate the fifth convolution result. The third convolution result is input into the convolutional layer at the first convolutional kernel scale of the multi-scale regression network to generate the sixth convolution result. The fourth convolution result is input into the convolutional layer at the first convolutional kernel scale of the multi-scale regression network to generate the seventh convolution result. The first convolution result, the fifth convolution result, the sixth convolution result, and the seventh convolution result are combined to generate combined feature information; The combined feature information is input into the batch normalization layer and the activation function layer to output the atmospheric delay-corrected image.

5. The method according to claim 1, wherein, The step of extracting atmospheric delay parameter data from the target meteorological data and performing data dimension processing on the target elevation data to generate elevation-processed data includes: For each raster data in the target elevation data, perform the following generation steps: Divide the target value by the radar wavelength to generate the first division result; Multiply the satellite distance corresponding to the grid data with the sine value corresponding to the radar wave incident angle to obtain the multiplication result; Divide the vertical baseline length corresponding to the raster data by the multiplication result to obtain the second division result; The first division result, the second division result, and the ground height corresponding to the raster data are multiplied together to obtain the processed raster data; The resulting processed raster dataset is identified as the elevation processing data.

6. The method according to claim 3, wherein, The multi-branch feature encoding extractor further includes: a first multi-head attention mechanism model, a second multi-head attention mechanism model, and a third multi-head attention mechanism model; and The step of generating the multi-channel feature information set based on the low-frequency channel feature information, the high-frequency channel feature information, and the auxiliary channel feature information includes: The low-frequency channel feature information and the high-frequency channel feature information are input into the first multi-head attention mechanism model to generate the first low-frequency channel feature information and the first high-frequency channel feature information; The high-frequency channel feature information and the auxiliary channel feature information are input into the second multi-head attention mechanism model to generate the first auxiliary channel feature information and the second high-frequency channel feature information; The low-frequency channel feature information, the high-frequency channel feature information, and the auxiliary channel feature information are input into the third multi-head attention mechanism model to generate the second auxiliary channel feature information, the second low-frequency channel feature information, and the third high-frequency channel feature information; The low-frequency channel feature information, the high-frequency channel feature information, the auxiliary channel feature information, the first low-frequency channel feature information, the first high-frequency channel feature information, the first auxiliary channel feature information, the second high-frequency channel feature information, the second auxiliary channel feature information, the second low-frequency channel feature information, and the third high-frequency channel feature information are determined as the multi-channel feature information set.

7. An InSAR atmospheric delay correction device, comprising: The acquisition unit is configured to acquire target images, target meteorological data, and target elevation data captured by the target synthetic aperture radar for the target area; The processing unit is configured to extract atmospheric delay parameter data from the target meteorological data and to perform data dimension processing on the target elevation data to generate elevation-processed data. An interference processing unit is configured to perform interference processing on the target image to generate an interferogram; The extraction unit is configured to extract high-frequency component features and low-frequency component features from the interferogram using each of at least one wavelet transform method, so as to generate high-frequency component feature information and low-frequency component feature information. The first fusion unit is configured to perform data interpolation fusion on the elevation processing data and the atmospheric delay parameter data to generate auxiliary matrix data with the same size as the matrix corresponding to the interferogram. The second fusion unit is configured to fuse at least one high-frequency component feature information to generate first fused feature information, and to fuse at least one low-frequency component feature information to generate second fused feature information. The generation unit is configured to input the first fused feature information, the second fused feature information, and the auxiliary matrix data into a pre-trained atmospheric delay correction neural network model to generate an atmospheric delay correction image for the target image.

8. An electronic device, comprising: One or more processors; Storage device, on which one or more programs are stored, When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1-6.

9. A computer-readable medium having a computer program stored thereon, wherein, When the program is executed by the processor, it implements the method as described in any one of claims 1-6.

Images