Accumulated snow coverage monitoring method based on RepFNet network

Abstract

The invention relates to the technical field of remote sensing image processing and meteorological monitoring, in particular to a RepFNet network-based accumulated snow coverage monitoring method, which comprises the following steps of: acquiring FY-4A image data, Landsat satellite image data and geographical auxiliary data; constructing a data set suitable for deep learning training; generating a high-resolution snow coverage image; the method comprises the following steps: constructing a RepFNet network model of an encoder-decoder architecture, and integrating an improved feature extraction module, a dynamic up-sampling module and an adaptive graph channel attention module; determining an optimal module combination and parameter configuration; performing model training by using an ADAM optimizer in combination with a dynamic learning rate strategy; the model performance is optimized by adjusting hyper-parameters and contrast experiments; and the RepFNet model after training is utilized to carry out accurate inversion of the snow coverage degree on a research area, and cloud removal processing is carried out in combination with a space-time filtering method, so that the accuracy and stability of snow monitoring are ensured.

Description

Technical Field

[0001] The invention relates to the technical field of remote sensing image processing and meteorological monitoring, and in particular to a snow cover monitoring method based on a RepFNet network. Background Art

[0002] Snow is an important physical element on the earth's surface with unique characteristics. It is widely distributed in different climate zones and has an important impact on the global environment and climate change research. Fractional snow cover (FSC) is calculated by calculating the percentage of snow within the pixels of remote sensing imagers and provides a more accurate estimate of the snow area than binary snow identification. Traditional snow monitoring methods mainly rely on ground observations and manual field detection, but these methods face many challenges in areas with high altitudes and complex terrain, such as high cost, low efficiency, and difficulty in covering large areas. With the development of remote sensing technology, its effective way to perceive large-scale snow information is also of greater significance for achieving more accurate snow cover.

[0003] In recent years, deep learning technology has made significant progress in the field of image processing and pattern recognition, providing new ideas for snow cover inversion. Deep learning models can automatically learn complex features in input data and have stronger feature extraction capabilities and robustness. However, most of the existing deep learning-based FSC inversion methods rely on traditional convolutional neural network architectures, which often cause information loss due to the increase in the number of network layers when processing high-resolution remote sensing images, affecting the inversion accuracy. In addition, when fusing multi-source data (such as satellite images and geographic elevation data), existing methods fail to give full play to the advantages of each data source, resulting in insufficient accuracy and reliability of the inversion results. Therefore, in terms of using remote sensing data for earth environment monitoring, the performance of deep learning models far exceeds that of traditional models, and can provide more accurate snow cover predictions in different regions and seasons.

[0004] Traditional snow cover monitoring methods face two major bottlenecks in complex mountainous terrain: first, most satellites have long revisit periods, making it difficult to capture rapid changes in snow cover (such as snowmelt and snowfall events), resulting in limited dynamic monitoring capabilities; in addition, the climate in mountainous areas is changeable, and single-phase data is susceptible to cloud contamination, resulting in data loss and inversion error accumulation. Therefore, a snow cover inversion method based on the RepFNet network and spatiotemporal filtering is developed, and daily snow monitoring products are generated by efficiently fusing multi-source data (including remote sensing images, meteorological data, and terrain data) and de-clouding processing. This has important practical significance for improving the real-time, accuracy, and robustness of snow monitoring, especially in complex terrain areas in mountainous areas. Summary of the invention

[0005] The technical problem to be solved by the present invention is to overcome the shortcomings of the existing technology in snow cover monitoring and cloud removal processing in complex mountainous terrain, and provide a snow cover monitoring method based on the RepFNet network. The method introduces high-temporal resolution FY-4A / AGRI meteorological satellite data, combined with the efficient feature extraction capability and spatiotemporal filtering strategy of the deep learning model to improve the accuracy and stability of snow cover inversion.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: A snow cover monitoring method based on a RepFNet network, the method comprising: S100, obtain FY-4A advanced geosynchronous radiation imager data in good weather, and simultaneously obtain Landsat satellite image data with a time difference of no more than 10 minutes to ensure the timeliness of the data, and perform band fusion and resampling operations on Landsat data to match the spatial resolution of FY-4A data; obtain and splice digital elevation model (DEM) data of the study area, and calculate slope (Slope) and slope aspect (Aspect) information based on DEM as auxiliary terrain factors for snow cover estimation; S200, spatially align the pre-processed FY-4A meteorological satellite data, Landsat data, and DEM-derived terrain information, and crop them into 64×64 pixel image blocks to construct a dataset suitable for deep learning training; S300, using the SNOMAP algorithm to identify snow cover in binary Landsat data, and combining the Normalized Difference Snow Index (NDSI) with the resolution of meteorological satellites to calculate snow cover, generating a high-resolution snow cover image as label data for the deep learning model; S400, using the PyTorch framework, build and implement the RepFNet network model with an encoder-decoder architecture. The model integrates an improved feature extraction module, a dynamic upsampling module (DySample), and an adaptive graph channel attention (AGCA) module to enhance the detection capability of snow areas and improve the inversion accuracy. S500, deeply optimize and adjust the parameters of the RepFNet network model, adjust key network components such as the feature extraction module and the attention mechanism module, and determine the best module combination and parameter configuration through a large number of experiments to improve the stability and generalization ability of the model; S600, design a new loss function suitable for FSC inversion, and use the ADAM optimizer combined with a dynamic learning rate strategy to train the model. Optimize model performance by repeatedly adjusting hyperparameters and conducting comparative experiments. Finally, use the trained RepFNet model to accurately invert snow cover in the study area, and combine it with the spatiotemporal filtering method to remove clouds, ensuring the accuracy and stability of snow monitoring.

[0007] Preferably, FY-4A / AGRI meteorological satellite data is obtained from the Fengyun satellite remote sensing data service network; land satellite data is obtained from the official website of the United States Geological Survey (USGS); and digital elevation data is obtained from the official website of the Geospatial Data Cloud. Preferably, the slope and aspect information is calculated based on the digital elevation model in S100, including: Using DEM data, slope and aspect data are calculated through Python scripts to generate terrain auxiliary data layers that are consistent with the FY-4A and Landsat data space, providing terrain feature support for subsequent model training.

[0008] Preferably, S200 includes: ENVI5.3 software was used to define the size of the clipping area as 64×64 pixels according to the study area, and a vector file was generated. FY-4A meteorological satellite data, Landsat land satellite data, and geographic auxiliary data (slope, slope aspect) were loaded respectively, and Landsat data was used as the reference data for spatial alignment, and each data was clipped in turn.

[0009] Preferably, generating a high-resolution snow cover image in S300 includes: The normalized difference snow index NDSI is determined according to the formula: Among them, Band2 is the green light band of Landsat8 data, and Band6 is the short-wave infrared band of Landsat8 data; The Landsat data was converted into a binary snow map by selecting the NDSI threshold and then resampled to ensure spatial consistency with the FY-4A data. The snow coverage within each pixel is calculated based on the number of snow pixels and the total number of pixels to generate a snow coverage map.

[0010] Preferably, the RepFNet network model in S400 includes: an improved feature extraction module, a dynamic upsampling module and an adaptive graph channel attention module; The network model constructed by the present invention performs feature extraction by improving the RepVGG module, proposes a dynamic residual weight driven by local variance, and realizes the adaptive adjustment of the branch ratio of this module. In order to gradually restore the high-dimensional features output by the feature extraction module to the same spatial resolution as the input data, the present invention also introduces a decoding module on this basis. The decoding module adopts a step-by-step upsampling method, and gradually restores the spatial information of the feature map through the alternating combination of convolutional layers and upsampling layers. The dynamic sampling module DySample is used for upsampling, and the upsampling strategy is adaptively adjusted. In addition, the AGCA module is introduced, which can adaptively adjust the weights of the feature channels and capture the global dependencies between features.

[0011] Preferably, the deep optimization and parameter adjustment in S500 include: The threshold of network iterations was set to 200 to ensure that the model converged within a limited number of iterations. The initial learning rate was set to 0.001, and the ADAM optimizer was used to update the model parameters. After every 10 iterations, the learning rate is dynamically adjusted with a decay coefficient of 0.8 to accelerate model convergence and prevent overfitting.

[0012] Preferably, in S600, the trained RepFNet model is used to accurately invert the snow cover of the study area, including: The model accuracy was verified using the test data set, and the evaluation indicators included the coefficient of determination, root mean square error, correlation coefficient, Kappa coefficient, and explained variance score; Among them, the coefficient of determination for: The root mean square error RMSE is: .

[0013] Correlation coefficient for: .

[0014] The Kappa coefficient is: .

[0015] The explained variance score EVS is: in, represents the true value, represents the predicted value; represents the mean of the true values, represents the mean of the predicted values, represents the observation accuracy, that is, the amount by which the predicted value matches the true value of the classification. represents random accuracy, which is the proportion of the model calculated based on the probability of random classification matching the actual category, and n represents the size of the sample; Among them, when the determination coefficient, EVS, correlation coefficient and Kappa value are closer to 1, the model fits the data better; when the RMSE is closer to 0, the model inversion accuracy is higher.

[0016] Preferably, S600 combines a spatiotemporal filtering method to perform cloud removal processing to ensure the accuracy and stability of snow monitoring, including: Deploy the trained RepFNet model to the study area data and output the snow cover inversion results; The data in the study area are subjected to multi-temporal declouding processing, and the changes in snow cover are monitored after declouding in combination with time series data to ensure the stability and timeliness of the inversion results.

[0017] The computer device described in the present invention includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the snow cover inversion method based on the RepFNet network are implemented.

[0018] The computer-readable storage medium of the present invention stores a computer program, and when the computer program is executed by a processor, the steps of the snow cover inversion method based on the RepFNet network are implemented. Compared with the prior art, the beneficial effects achieved by the present invention are: The present invention does not require additional hardware equipment. It combines FY-4A / AGRI meteorological satellite data, Landsat land satellite data and terrain auxiliary data (slope, aspect, DEM) to achieve deep fusion of multi-source data and improve the accuracy of snow cover inversion under complex terrain. The RepFNet network is constructed by improving the RepVGG module, and the dynamic residual weight strategy driven by local variance is integrated to improve the robustness and adaptability of feature extraction, especially in variable climate conditions and complex backgrounds. The dynamic upsampling module DySample and adaptive global channel attention are introduced to improve the robustness and adaptability of feature extraction. The force module AGCA not only enhances the recognition accuracy of snow edge areas and fragmented snow areas, but also can adaptively adjust the upsampling strategy and feature channel weights to capture the spatial details and global dependencies of the snow area, avoiding the problems of over-smoothing and boundary blur in traditional methods; combined with declouding processing and high temporal resolution reconstruction strategy, through multi-temporal declouding technology, it effectively solves the problem of data missing caused by cloud occlusion, ensures the continuity and integrity of the inversion results in space and time, and realizes large-scale, real-time and high-precision snow cover inversion, providing more reliable data support for snow monitoring, climate research and disaster prevention and mitigation. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings: Figure 1 It is a schematic diagram of the process of the present invention.

[0020] Figure 2 It is the preprocessed FY-4A / AGRI image data in the embodiment of the present invention.

[0021] Figure 3 It is the landsat8 data in the embodiment of the present invention.

[0022] Figure 4 It is the image data after the DEM elevation data is spliced ​​in the embodiment of the present invention.

[0023] Figure 5 Schematic diagram of a label in an embodiment of the present invention.

[0024] Figure 6 Schematic diagram of inputting data in an embodiment of the present invention.

[0025] Figure 7 Schematic diagram of the structure of the snow cover estimation model in an embodiment of the present invention.

[0026] Figure 8 Schematic diagram of a feature extraction network in an embodiment of the present invention.

[0027] Fig. 9 Schematic diagram of the structure of an upsampling module in an embodiment of the present invention.

[0028] Fig.10 Schematic diagram of the structure of the attention mechanism module in an embodiment of the present invention.

[0029] Fig.11 Schematic diagram of the snow cover inversion result and cloud removal result with high temporal resolution in an embodiment of the present invention.

[0030] Fig.12 This is a schematic diagram of a daily product that is finally successfully constructed in an embodiment of the present invention.

[0031] Fig.13 It is a comprehensive and clear picture of the final daily product in the embodiment of the present invention. DETAILED DESCRIPTION

[0032] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0033] See also Figure 1-Figure 13 , the present invention provides a technical solution:

[0034] Example 1: Figure 1 As shown, the present invention designs a snow cover inversion method based on the RepFNet network, and the method specifically includes the following steps: In S100: FY-4A / AGRI image data with good weather conditions is selected from the Fengyun satellite remote sensing data service network. The present invention uses bands 1 to 7 of the FY4A / AGRI satellite image, performs radiometric calibration and geometric correction on the FY-4A image data, and obtains the following Figure 2 Image data of the required area can be obtained by clipping the remote sensing data according to the boundary of the vector data; through the GEE platform, land satellite image data with a time of no more than 10 minutes from the FY-4A / AGRI image data can be selected, such as Figure 3 As shown in the figure, the bands of the study area were cropped and fused, including the B3, B4, B5 and B6 bands, to improve the spatial resolution and information content of the image. The Landsat image data after band fusion was resampled to 2000m resolution to make it consistent with the resolution of FY-4A / AGRI image data.

[0035] In S200: the geographic elevation data of the study area is obtained and spliced ​​to ensure that the data within the study area is complete. The final elevation data is as follows: Figure 4 shown.

[0036] By manually drawing multiple 64×64 pixel rectangular area vector files on the Landsat8 data, we used them as templates for subsequent cropping. We loaded the preprocessed FY-4A / AGRI meteorological satellite data, Landsat8 land satellite data, and geographic auxiliary data, respectively, to ensure that the spatial resolution and projection information of all data were consistent. We used georeferencing to spatially align the data used, and used the cropping template to create a 2000m resolution dataset, and then performed random horizontal and vertical operations on it to expand the dataset. The labels and input data are as follows: Figure 5 , Figure 6As shown. In S300: the processed land satellite image data is processed using the SNOMAP algorithm, an NDSI mask is generated according to preset conditions (the NDSI threshold selected in the present invention is 0.29), valid data is screened, and a binary snow map is generated to distinguish between snow-covered and non-snow-covered areas. Since the reflection characteristics of snow and water bodies in the short infrared band and the visible band are very similar, a determination factor is introduced to determine whether the reflectivity of the fifth band is greater than or equal to 0.11 to eliminate the interference of water bodies on snow recognition and adjust the spatial resolution of the binary snow map to be consistent with the FY-4A meteorological satellite data. Subsequently, the snow cover is calculated for the binary snow map, and the snow cover (FSC) calculation formula is as follows: .

[0037] in, Represents the number of snow pixels in each computational unit, Indicates the total number of pixels in the calculation unit. After the snow cover map is generated by calculation, verify whether the value range of the FSC map is between 0% and 100%, and check whether there are any outliers.

[0038] In S400: The Pytorch framework is used to build the RepFNet network model for snow cover inversion tasks. The overall network structure is shown in the figure below: Figure 7 The present invention uses the improved RepVGG module to effectively extract the texture and spectral features of the snow-covered area, and uses the residual connection to enhance the training stability and performance of the model, as shown in Figure 8 As shown in the figure, a decoding module is added based on this structure, and a dynamic upsampling module DySample is introduced for upsampling, and the AGCA module is used to improve the feature expression capability.

[0039] This network is inspired by the residual structure of the ResNet network and is trained with fixed-ratio identity branches and 1×1 convolution branches. However, the texture complexity of different areas in the snow scene varies significantly, and fixed weights cannot adaptively extract key features. Therefore, the present invention introduces dynamic residual weights in the RepVGG module of feature extraction, and dynamically adjusts the branch weights according to the local texture complexity of the input feature map. The formula is as follows: .

[0040] .

[0041] Var(Fin) represents the local lengthening of the feature map Fin of the input data (sliding window 3×3), which represents the complexity of regional texture: when the texture is complex like the edge of snow, it is a high variance area. Close to 1, dominated by the identity branch, retaining more original feature details; flat areas such as continuous snow layers are low variance areas. Close to 0, dominated by 1×1 convolution, enhancing feature extraction capabilities. The initial value of is set to 0.5 and is automatically optimized after training.

[0042] At the same time, we choose to use two-dimensional convolution to extract spatial features from the image. Before the true value of the snow cover label is sent to the encoder, a convolution with a stride of 2 is used for downsampling. The next block uses a convolution with a stride of 1 to extract features and compress the feature map into a low-dimensional representation. Before outputting the prediction map, we use the dynamic upsampling module DySample to perform an upsampling operation on the output of the RepVGG decoder, so that it and the transposed convolution layer gradually restore the spatial resolution of the feature map, and restore the encoded feature information to the spatial resolution of the original image to generate high-quality inversion results. For DySample, a dynamic range factor is introduced , to increase the flexibility of the offset. Finally, the output feature map is obtained by bilinear interpolation , where X represents the input and S represents the sampled output graph; .

[0043] After convolution by RepVGG block, some fragmented snow areas cannot be captured accurately. By adding an adaptive graph channel attention (AGCA) module, local and global context information can be aggregated to improve the expressiveness of features.

[0044] like Fig. 9 As shown, the module mainly consists of two parts.

[0045] The first part is the feature mapping layer, which converts the feature map dimension to 1×1×C (C represents the number of channels of the image), and also represents the number of feature vertices in the AGCA module.

[0046] Using a linear embedding function As a feature mapping layer, it focuses on the characteristics of important snow areas and fragmented snow, so that the entire model can more effectively capture the key information of snow cover. The module fuses this information to produce improved feature representations. These feature maps can better characterize the complex structure and semantic information of the input data. Where W is a weight matrix to be learned through 1×1 convolution.

[0047] The second part is to dynamically adjust the feature weights through the adaptive image convolution module (AGCM), and dynamically adjust the feature weights through its designed attention mechanism, and finally map the weights back to the original feature map as channel weights. The AGCA module also uses a bottleneck structure to make its feature vertices more sensitive to the feature redundancy of some snow accumulation. The introduction of the AGCA module enables the model to more adaptively focus on different areas and features of snow accumulation, thereby improving the generalization ability of the model.

[0048] In S500: In order to improve the accuracy of the model in processing FSC prediction, the present invention designs a new loss function suitable for this task, and this time uses a combination of two functions: and ; in It is a robust loss function (Huber Loss). Nowadays, many regression tasks use mean square error (MSE) as the loss function. Compared with MSE, Huber Loss is not easily affected by outliers. At the same time, compared with mean absolute error (MAE), Huber Loss can provide a smooth optimization process, which is conducive to the convergence of the optimization algorithm. Therefore, Huber Loss can take into account both convergence speed and outlier processing capabilities. The specific formula is as follows: .

[0049] in, is a hyperparameter (in this paper, = 1.3), which is used to control the sensitivity to errors. x represents the true label image of FSC, and y represents the label prediction image of FSC; is the gradient difference loss (GDL), which is a strategy that can sharpen the edges of images. In the generated loss function, the gradient difference of the predicted image is directly penalized, which can effectively alleviate the phenomenon of image edge blur. It is shown in the following formula: .

[0050] in, is an integer. The loss function used in this study is composed of Huber Loss and GDL, as shown in the following formula: .

[0051] in, and is the weight of the loss function, satisfying .at this time Combining the advantages of both, it is introduced into the RepFNet model, so as to try to balance different loss characteristics in the inversion of snow cover. In order to give the best performance of this loss function in the FSC regression training stage, this study adjusts the loss function in the training stage. and This study selected multiple weight combinations and evaluated the effects of different weights by comparing the performance of the model (as shown in the table below). The final selection results are shown in This combination is the best combination. In S600: Adjust the parameters of the RepFNet network model, including learning rate, batch size, number of iterations, number of weight decay, etc.

[0052] The ADAM optimizer is selected for model optimization, and the convergence speed and accuracy of the model are improved by dynamically adjusting the learning rate. During the model training process, by continuously optimizing the parameters, the present invention can find the most suitable parameter combination, thereby constructing the best snow cover inversion model. Before training the model, in order to ensure that the data of each channel are at the same order of magnitude and avoid the problem of gradient explosion during the training process, the present invention performs normalization preprocessing on the input data. In this process, we use the maximum and minimum normalization method to ensure the stability of the data and the convergence of the model.

[0053] Among them, the formula for normalization of maximum and minimum values ​​is as follows: .

[0054] Among them, x is the original data, and are the minimum and maximum values ​​in the data channel, respectively. is the normalized data.

[0055] In the implementation process of the present invention, through the analysis of a large amount of experimental data and the continuous optimization of model training, the model parameter configuration most suitable for this example was finally determined. These parameters have been adjusted many times to ensure that the model achieves the best performance in the snow cover inversion task. The specific parameter values ​​are as follows: the number of training iterations (epoch) is set to 200 times, the initial learning rate (learning_rate) is 0.001, and the learning rate attenuation coefficient (gamma) is 0.8. The indicators for model accuracy verification include the determination coefficient ( ), root mean square error (RMSE), correlation coefficient, Kappa coefficient and explained variance fraction (EVS). The accuracy of the optimal model is: R2=0.6763, MSE=0.0956, Correlation=0.8539, EVS=0.6602, Kappa=0.4728, which proves the effectiveness of the present invention. Finally, multiple FY-4A / AGRI image data of the required area per day are downloaded, and the snow cover map of the required area is inverted through the RepFNet model. By inverting meteorological image data of multiple hourly time periods in a day, intraday fusion of multi-temporal remote sensing images is performed to achieve cloud removal of the final inversion result. The flow chart is shown in the figure. Fig.11 As shown. The data of the ith time is recorded as Xi(x,y), where (x,y) represents the pixel position in the image. Clouds usually show high reflectivity or high brightness, and other ground objects (such as snow or land) are relatively stable. The reflectivity of clouds often changes between multiple time periods. The present invention defines the snow accumulation judgment threshold Ts as 0.2. When Xi(x,y) is greater than or equal to 0.2, the pixel is judged as snow accumulation, recorded as Si(x,y)=1, otherwise it is 0.

[0056] For pixels at the same location, the frequency of snow accumulation in a day is calculated as F(x,y): .

[0057] When F(x,y) is greater than 0.7, the pixel is considered to be a stable snow pixel SN(x,y), otherwise it is marked as a suspicious cloud area.

[0058] At the same time, define the number of pixels within 3×3 pixels centered on the pixel (x, y) as N(x, y), and count the proportion of snow pixels in the area P(x, y): .

[0059] If the ratio exceeds the spatial consistency threshold Tp (the threshold is set to 0.6 in the present invention), the central pixel is considered to have snow accumulation consistency in space and is judged to be snow accumulation.

[0060] After the snow accumulation is determined, the suspected cloud area is repaired using a multi-time fusion strategy, and valid observations from other times of the same day are selected to fill in the gaps in a weighted average manner according to the proximity of the observation time to the government. Fig.12 As shown in Figure 1, the inversion results of 7 time points in one day were finally declouded (as shown in h). Following the above steps, the daily product was successfully constructed, achieving high-precision monitoring of the dynamic changes of daily snow cover, providing strong support for comprehensively and clearly presenting the evolution characteristics of snow cover in time series, as shown in Figure 1. Fig.13 shown.

[0061] Embodiment 2: This embodiment provides a snow cover inversion device based on the RepFNet network, which is deployed on an edge device or in the cloud, including: The data acquisition module is used to acquire remote sensing image data (such as satellite images, drone images, etc.), pre-process the data, and construct a snow cover inversion dataset; A model building module is used to build a snow cover inversion model based on the RepFNet network. The model training process and the process of obtaining the true value of snow cover are the same as those in Example 1. The snow cover inversion module is used to obtain real-time remote sensing image data and call the trained RepFNet network model to obtain the current snow cover inversion result.

[0062] Embodiment 3: This embodiment provides a non-transitory computer-readable storage medium having computer instructions stored thereon, which enable a computer to execute the snow cover inversion method based on the RepFNet network in Embodiment 1.

[0063] Embodiment 4: This embodiment provides an electronic device, which may include a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the snow cover inversion method based on the RepFNet network when executing the computer program. The memory may include a memory, such as a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk memory, etc. The processor, the network interface, and the memory are interconnected through an internal bus, and the internal bus may be an industrial standard architecture bus, a peripheral component interconnection standard bus, an extended industrial standard structure bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. The memory is used to store programs, and specifically, the program may include a program code, and the program code includes computer operation instructions. The memory may include a memory and a non-volatile memory, and provide instructions and data to the processor.

[0064] In the computer device, when the processor executes the computer program, the following steps are specifically implemented: The remote sensing image data is acquired through the data acquisition module, and the data is preprocessed to construct a snow cover inversion dataset; The AGCA module and Dysample module are used to enhance data features and dynamically sample data to optimize data distribution. Construct and train a snow cover inversion model based on the RepFNet network; Call the trained RepFNet network model to invert the snow cover of real-time remote sensing image data and output the inversion results.

[0065] Embodiment 5: This embodiment provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the snow cover inversion method based on the RepFNet network in Embodiment 1.

[0066] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0067] Based on this understanding, the above technical solution can essentially or contribute to the prior art in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or certain parts of the embodiments.

[0068] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A snow cover monitoring method based on RepFNet network, characterized in that: The method comprises: S100, obtaining FY-4A image data and Landsat satellite image data that is no more than 10 minutes old from the image data, and performing band fusion and resampling operations on the Landsat satellite image data; obtaining and splicing digital elevation model data of the study area, and calculating slope and aspect information based on the digital elevation model as geographic auxiliary data for snow cover estimation; S200, spatially registering the preprocessed FY-4A image data, Landsat data, and geographic auxiliary data with Landsat data as the reference data, and cropping them into 64×64 pixel image blocks to construct a data set suitable for deep learning training; S300, using the SNOMAP algorithm to identify snow cover in binary Landsat data, and combining the normalized difference snow index with the resolution of meteorological satellites to calculate the snow cover, generating a high-resolution snow cover image as label data for the deep learning model; S400. Using the PyTorch framework, build and implement the RepFNet network model with encoder-decoder architecture; the model integrates an improved feature extraction module, a dynamic upsampling module, and an adaptive graph channel attention module; S500, determining the best module combination and parameter configuration based on the deep optimization and parameter adjustment of the RepFNet network model; S600, design a new loss function suitable for FSC inversion, and use the ADAM optimizer combined with a dynamic learning rate strategy to train the model; optimize the model performance by adjusting hyperparameters and conducting comparative experiments; use the trained RepFNet model to accurately invert the snow cover in the study area, and combine the spatiotemporal filtering method for cloud removal to ensure the accuracy and stability of snow monitoring.

2. A snow cover monitoring method based on RepFNet network as claimed in claim 1, characterized in that: The step S100 calculates the slope and aspect information based on the digital elevation model, including: Using digital elevation model data, slope and aspect data are calculated through Python scripts to generate terrain auxiliary data layers that are consistent with the FY-4A and Landsat data space, providing terrain feature support for subsequent model training.

3. A snow cover monitoring method based on RepFNet network as claimed in claim 1, characterized in that: The step S300 generates a high-resolution snow cover image, including: The normalized difference snow index NDSI is determined according to the formula: Among them, Band2 is the green light band of Landsat8 data, and Band6 is the short-wave infrared band of Landsat8 data; The Landsat data was converted into a binary snow map by selecting the NDSI threshold and then resampled to ensure spatial consistency with the FY-4A data. The snow coverage within each pixel is calculated based on the number of snow pixels and the total number of pixels to generate a snow coverage map.

4. A snow cover monitoring method based on RepFNet network as claimed in claim 1, characterized in that: The RepFNet network model in S400 includes: an improved feature extraction module, a dynamic upsampling module and an adaptive graph channel attention module; The improved feature extraction module is used to propose a local variance driven dynamic residual weight to achieve adaptive adjustment of the module branch ratio; The dynamic upsampling module is used to gradually restore the high-dimensional features output by the feature extraction module to the same spatial resolution as the input data; it adopts a step-by-step upsampling method, through the alternating combination of convolutional layers and upsampling layers, to gradually restore the spatial information of the feature map; The adaptive graph channel attention module is used to adaptively adjust the weights of feature channels while capturing the global dependencies between features.

5. The snow cover monitoring method based on RepFNet network as claimed in claim 1, characterized in that: The in-depth optimization and parameter adjustment in S500 include: The threshold of network iterations was set to 200 to ensure that the model converged within a limited number of iterations. The initial learning rate was set to 0.001, and the ADAM optimizer was used to update the model parameters. After every 10 iterations, the learning rate is dynamically adjusted with a decay coefficient of 0.8 to accelerate model convergence and prevent overfitting.

6. A snow cover monitoring method based on RepFNet network as claimed in claim 1, characterized in that: In S600, the trained RepFNet model is used to accurately invert the snow cover of the study area, including: The model accuracy was verified using the test data set, and the evaluation indicators included the coefficient of determination, root mean square error, correlation coefficient, Kappa coefficient, and explained variance score; Among them, the coefficient of determination for: The root mean square error RMSE is: Correlation coefficient for: The Kappa coefficient is: The explained variance score EVS is: in, represents the true value, represents the predicted value; represents the mean of the true values, represents the mean of the predicted values, represents the observation accuracy, that is, the amount by which the predicted value matches the true value of the classification. represents random accuracy, which is the proportion of the model calculated based on the probability of random classification matching the actual category, and n represents the size of the sample; Among them, when the determination coefficient, EVS, correlation coefficient and Kappa value are closer to 1, the model fits the data better; when the RMSE is closer to 0, the model inversion accuracy is higher.

7. The snow cover monitoring method based on RepFNet network as claimed in claim 1, characterized in that: The S600 combines the spatiotemporal filtering method to perform cloud removal processing to ensure the accuracy and stability of snow monitoring, including: The trained RepFNet model is deployed on the data of the study area to output the snow cover inversion results. The data in the study area is subjected to multi-temporal declouding, and the time series data is combined with declouding to monitor the changes in snow cover and ensure the stability and timeliness of the inversion results.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps in a snow cover monitoring method based on a RepFNet network as described in any one of claims 1 to 7 are implemented.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the program, the steps in the snow cover monitoring method based on the RepFNet network as described in any one of claims 1-7 are implemented.

Images