Multiband radar data spatial downscaling method based on data fusion

Abstract

The invention provides a multi-band radar data space downscaling method based on data fusion. The method comprises the following steps: step 1, acquiring S-band and X-band observation data from a weather radar; step 2, carrying out data preprocessing on the obtained observation data; step 3, constructing a spatial downscaling model; 4, performing training verification and testing on the spatial downscaling model through the observation data after data preprocessing; step 5, according to the trained spatial downscaling model, performing spatial downscaling of the S-band observation data; and step 6, judging whether the generated data is close to a true value or not. By means of the deep learning neural network, the model can extract and fuse the characteristics of the S-band observation data and the X-band observation data, the original low-resolution S-band data is converted into high-resolution data, the requirement for high resolution of radar data in practical application is met, and dependence on hardware facilities is reduced.

Description

Technical Field

[0001] The present invention relates to the fields of weather radar and spatial downscaling, and in particular to a multi-band radar data spatial downscaling method based on data fusion. Background Art

[0002] Weather radar is a key tool for monitoring and analyzing atmospheric phenomena in meteorology. With the development of technology, spatial downscaling technology based on deep learning is gradually being applied to the reconstruction process of weather radar data. Currently, many research teams have proposed innovative spatial downscaling methods based on super-resolution reconstruction. However, these existing methods all have an inherent defect, that is, the low-resolution data they need are all obtained by a specific processing flow of the corresponding high-resolution data. This data acquisition method makes low-resolution data dependent on high-resolution data at the source, which may cause doubts about data authenticity, poor data reliability, and potential data consistency issues in practical applications. This ultimately affects the universality and effectiveness of such methods in different application scenarios, and it is difficult to fully meet the needs of actual meteorological monitoring and analysis for high-quality and high-reliability radar data.

[0003] Currently, there is no algorithm that focuses on feature fusion processing of S-band observation data and X-band observation data to construct a spatial downscaling model. Summary of the invention

[0004] Purpose of the invention: The technical problem to be solved by the present invention is to provide a multi-band radar data spatial downscaling method based on data fusion in view of the shortcomings of the prior art. The method can convert the original low-resolution S-band data into high-resolution data, meet the demand for high-resolution radar data in practical applications, and reduce the dependence on hardware facilities.

[0005] The method of the present invention comprises the following steps: Step 1, obtaining S-band and X-band observation data from weather radar; Step 2, preprocessing the acquired observation data; Step 3, construct a spatial downscaling model; Step 4, training, verifying and testing the spatial downscaling model through the observation data after data preprocessing; Step 5, perform spatial downscaling of the S-band observation data according to the trained spatial downscaling model; Step 6: Determine whether the generated data is close to the true value.

[0006] Step 2 includes: Step 2.1, time-space matching of S-band and X-band observation data: According to the time identification information contained in the original observation file name, the X-band observation data and the S-band radar are time-calibrated so that the X-band observation data and the S-band radar are aligned in the time dimension; According to the latitude and longitude information corresponding to different stations, the S-band observation data is clipped according to the latitude and longitude values; Step 2.2, rotate and align the X-band observation data: By analyzing the historical observation data of the X-band, the current angle deviation value of the X-band is determined; Based on the angle deviation value, the radar scanning center is used as the rotation reference point, and a rotation matrix algorithm based on the principle of trigonometric function is used to transform the coordinate system of the X-band observation data; Step 2.3, interpolate the X-band observation data to fill in the missing data: Based on the distribution characteristics of the valid data area acquired by the X-band radar, the data range of invalid values ​​in the data is calculated to define the scanning blind area range and boundary conditions where data is missing; Select an interpolation algorithm, use the valid data points around the scanning blind area as the reference data source, perform data estimation according to the established interpolation rules, and fill the estimated data into the corresponding missing positions; Step 2.4, remove low-quality data: perform quality assessment on the X-band observation data and the S-band observation data, and delete the low-quality radar data from the X-band observation data and the S-band observation data; use the X-band observation data as high-resolution input data and the S-band observation data as low-resolution input data.

[0007] Step 3 includes: Step 3.1, construct two independent feature extraction modules to process low-resolution input data and downsampled high-resolution input data respectively; Multi-scale feature extraction for low-resolution input data: The S-band observation data is input into the dilated multi-scale convolutional network, and flows through the small receptive field convolution layer, the medium receptive field convolution layer and the large receptive field convolution layer in sequence to extract multi-scale features and generate a low-resolution LR feature map; Downsampling and multi-scale feature extraction are performed on high-resolution input data: First, downsampling is performed on the X-band observation data to reduce the spatial resolution to be consistent with the S-band observation data, and downsampled high-resolution input data HR Down Input is generated to simulate the data distribution under low-resolution input conditions; secondly, the high-resolution input data HR Down Input is input into the same hole multi-scale convolutional network structure as the low-resolution branch, and flows through the small receptive field convolution layer, the medium receptive field convolution layer and the large receptive field convolution layer in sequence to extract multi-scale features and generate a downsampled high-resolution HR Down feature map. Through the symmetrical feature extraction structure, the model learns the common features of low-resolution and high-resolution data; Step 3.2, build data fusion module: The data fusion module deeply fuses the features extracted from the two branches to generate a comprehensive feature map: first, a channel splicing operation is performed to splice the low-resolution LR feature map after multi-scale feature extraction with the downsampled high-resolution HR Down feature map according to the channel dimension. The spliced ​​feature map passes through the convolution layer, batch normalization layer and rectified linear unit in turn to achieve deep fusion of the features from the low-resolution LR and downsampled high-resolution HR Down data; The convolution layer in the data fusion module is a two-dimensional convolution layer, whose input is the spliced ​​128-channel feature map, and the convolution operation is performed by sliding a 3×3 convolution kernel on the feature map, and finally outputs a 64-channel feature map; the batch normalization layer is a two-dimensional batch normalization layer, which is used to normalize the data to a distribution with a mean of 0 and a variance of 1; the rectified linear unit is a ReLU activation function, which is used to perform nonlinear transformation on the batch-normalized feature map; the ReLU activation function is: , in is the input value after being processed by the batch normalization layer, is the output value after the ReLU activation function. If Greater than 0, output equal ;if Less than or equal to 0, output is equal to 0; Step 3.3, construct an upsampling module: the upsampling module includes a channel attention module, two residual dense blocks, a convolutional layer and bicubic interpolation upsampling; The channel attention module, two residual dense blocks, a convolutional layer for expanding the number of channels, bicubic interpolation upsampling, and a convolutional layer for adjusting the number of channels are connected in sequence: first, the channel attention module, followed by two residual dense blocks, and then the convolutional layer for expanding the number of channels, and then the bicubic interpolation function F.interpolate is used for upsampling to complete the conversion from low-resolution input to high-resolution output; finally, the convolutional layer for adjusting the number of channels is used to output the final upsampling result with the same number of channels as the input image; Among them, in the channel attention module, the spliced ​​feature map output by the data fusion module is first received; then the spliced ​​feature map is adaptively averaged and adaptively max-pooled respectively; then the results of average pooling and max-pooling are added together and input into a fully connected network consisting of two linear layers, ReLU activation function and Sigmoid function; then the shape of the attention weight is adjusted to the same size as the spliced ​​feature map; finally, the spliced ​​feature map is element-wise multiplied with the attention weight to obtain the feature map adjusted by the channel attention mechanism; The residual dense block performs the following operations: Step a1, during forward propagation, receive the feature map f adjusted by the channel attention mechanism; first, perform a convolution operation on the feature map f, and then activate it through the LeakyReLU function to obtain the feature map f1; then, concatenate the feature map f and the feature map f1 in the channel dimension, perform convolution and activate it through the LeakyReLU function to obtain the feature map f2; Step a2: According to step a1, the obtained feature maps f1 and f2 are continuously concatenated in the channel dimension and input into the convolution layer in the residual dense block to obtain feature maps f3 and f4 in turn; finally, all feature maps f1, f2, f3 and f4 are concatenated in the channel dimension and convolved to obtain feature map f5, and feature map f5 is multiplied by 0.2 and added to feature map f to obtain the output of the residual dense block; The formula for upsampling the bicubic interpolation function F.interpolate is: , in The function represents the use of the bicubic interpolation method to fused feature maps processed by the convolution layer with an expanded number of channels. Upsample to the specified size , and obtain a high-resolution feature map , complete the conversion from low-resolution input to high-resolution output; Indicates the height of the high-resolution image, Indicates the width of the high-resolution image; Step 3.4, modify the loss function: introduce height weight adjustment loss, perform linear mapping based on the input height information, associate height with weight, and introduce Sobel gradient loss to enhance edge information, including: The weight loss function formula is: , , in, represents the total loss, represents L1 loss, represents the Sobel gradient loss, represents the weight of Sobel gradient loss, n represents the number of samples, represents the total loss of the i-th sample, represents the height loss corresponding to the i-th sample, where i=1,2,…,n, represents the weighted loss; By sequentially constructing the feature extraction module, data fusion module, and upsampling module, we finally obtained the spatial downscaling model.

[0008] Step 4 includes: Step 4.1, divide the data set: The S-band and X-band observation data after data preprocessing are used as data pairs and divided into training set, validation set and test set according to proportion; the S-band observation data are used as low-resolution data, and the X-band observation data are used as high-resolution radar data, and the low-resolution data and high-resolution data correspond one to one; Step 4.2: Train the spatial downscaling model based on the divided data set: In the training phase, m S-band observation data in the training set are input. and m X-band observation data , the output prediction result is the image after model space downscaling, represents the mth S-band observation data, Represents the mth X-band observation data, where m is the number of data in the training set.

[0009] Step 4.2 also includes: updating the iterative weight parameters according to the weighted loss function of the spatial downscaling model through the back propagation algorithm, and performing cyclic iterative training until the model converges.

[0010] In step 5, the S-band observation data is spatially downscaled according to the trained spatial downscaling model; the data generated by the spatial downscaling model is visualized and compared with the actual X-band observation data.

[0011] In step 6, the evaluation index structural similarity SSIM is used to determine whether the generated data is close to the true value. The formula is: , Among them, the brightness comparison function , contrast comparison function and structure comparison functions The calculation formulas are: , , , in represents X-band observation data, represents the S-band observation data, and They are The mean and The mean of and They are The standard deviation and The standard deviation of yes and The covariance of , and is a constant, is the weight coefficient, when , it means that when calculating SSIM, the brightness comparison function , contrast comparison function and structure comparison functions The weight of the impact on the final result is the same. When the two images are the same, .

[0012] The present invention also provides an electronic device, comprising a processor and a memory, wherein the memory stores program code, and when the program code is executed by the processor, the processor executes the steps of the described method.

[0013] The present invention also provides a storage medium storing a computer program or instruction. When the computer program or instruction is run on a computer, the steps of the method described are executed.

[0014] Compared with the prior art, the present invention has the following beneficial effects: (1) The present invention accurately calibrates the S-band and X-band observation data to avoid data problems caused by time and space differences; at the same time, it comprehensively improves the quality of X-band observation data, which is superior to conventional simple processing methods. (2) The present invention adopts hollow multi-scale convolution to break through the limitation of a single scale and comprehensively mine data features from multiple dimensions; it innovatively integrates multi-scale features to achieve deep fusion and provide rich information for subsequent upsampling. (3) The present invention can use deep learning neural networks to allow the model to extract and integrate the features of S-band observation data and X-band observation data, and convert the original low-resolution S-band data into high-resolution data, which not only meets the demand for high-resolution radar data in practical applications, but also reduces the dependence on hardware facilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a flow chart of the present invention.

[0016] Figure 2 This is a data preprocessing flow chart of the present invention.

[0017] Figure 3 Schematic diagram of the structure of the spatial downscaling model of the present invention.

[0018] Figure 4 It is a structural schematic diagram of the feature extraction module of the present invention.

[0019] Figure 5 This is a structural diagram of the data fusion module of the present invention.

[0020] Figure 6 This is a training, validation and testing idea diagram for the spatial downscaling model of the present invention.

[0021] Figure 7 This is a rendering of the effect of the multi-band radar data spatial downscaling method based on data fusion of the present invention. DETAILED DESCRIPTION

[0022] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments, and the above and / or other advantages of the present invention will become more clear.

[0023] The embodiment of the present invention proposes a multi-band radar data spatial downscaling method based on data fusion, such as Figure 1 As shown, the following steps are included: Step 1: Obtain S-band and X-band observation data from weather radar: S-band and X-band observation data are obtained from weather radar. The S-band observation data are CAPPI products of networked contour reflectivity factors, and the X-band observation data are CAPPI products of phased array radar contour reflectivity factors. The spatial resolution of the X-band observation data is 0.0025, and the spatial resolution of the S-band observation data is 0.01. The time resolution is 6 minutes. The data altitude ranges from 2000 meters to 6000 meters, and is divided into altitude layers of 500 meters.

[0024] Step 2: preprocess the acquired observation data. The data processing flow chart is as follows: Figure 2 As shown, the specific steps include: Step 2.1, time-space matching of S-band and X-band observation data: Firstly, according to the time identification information contained in the original observation file name, the X-band observation data and the S-band radar are time-calibrated so that the X-band observation data and the S-band radar are aligned in the time dimension. Secondly, since the S-band radar is a networked CAPPI product, it is necessary to accurately clip the S-band observation data based on the latitude and longitude information corresponding to different stations to ensure its spatial consistency with the X-band observation data; Step 2.2, rotation and alignment of X-band observation data: Given that X-band radars have a certain degree of scanning angle deviation, this deviation will have an adverse effect on the precise alignment between different radar data. It is necessary to perform precise rotation adjustment operations on the X-band observation data to make its scanning angle consistent with the angle reference of the S-band observation data. Firstly, the current angle deviation value is determined by analyzing the historical scanning data of the X-band radar. Secondly, based on the determined deviation value, the radar scanning center is used as the rotation reference point, and the rotation matrix algorithm based on the principle of trigonometric function is used to transform the coordinate system of the X-band observation data. Step 2.3, interpolate and fill in missing data for X-band radar: In view of the problem of missing data in some areas due to scanning characteristics in X-band observation data, it is necessary to interpolate and fill in the missing data of X-band radar to ensure the integrity and continuity of the data; Firstly, based on the distribution characteristics of the valid data area acquired by the X-band radar, the data range of invalid values ​​in the data is calculated to accurately define the scanning blind area range and boundary conditions where data is missing. Secondly, select a suitable interpolation algorithm, such as an interpolation algorithm based on weighted average of adjacent points, a bilinear interpolation algorithm, a spline interpolation algorithm, or a Kriging interpolation algorithm, etc., use the valid data points around the scanning blind area as the reference data source, perform data estimation according to the established interpolation rules, and fill the estimated data into the corresponding missing positions; Step 2.4, remove low-quality data: The quality of X-band observation data and S-band observation data is evaluated, and the radar data with a small proportion of valid data is identified through the pre-set valid data amount threshold judgment standard. For example, if the proportion of zero values ​​in a certain type of data exceeds the threshold of 30%, then this type of data will be judged as data with a small proportion of valid data, that is, low-quality data; the low-quality radar data will be deleted from the X-band observation data and S-band observation data; the X-band observation data is used as high-resolution input data, and the S-band observation data is used as low-resolution input data.

[0025] Step 3: Construct a spatial downscaling model. The overall structure of the spatial downscaling model is shown in the figure below: Figure 3 As shown, it includes feature extraction module, data fusion module and upsampling module: Step 3.1, build two independent feature extraction modules to process low-resolution input data LowResolution Input (LR Input) and downsampled high-resolution input data High Resolution DownInput (HR Down Input) respectively; the specific structure of the module is as follows Figure 4 As shown: Multi-scale feature extraction for low-resolution input data (LR Input): The S-band observation data is input into the dilated multi-scale convolutional network, and flows through the small receptive field convolution layer (dilation rate 1, convolution kernel 3×3), the medium receptive field convolution layer (dilation rate 2, convolution kernel 3×3) and the large receptive field convolution layer (dilation rate 4, convolution kernel 3×3) in sequence to extract multi-scale features and generate a low-resolution LR feature map; The high-resolution input data (HR Input) is downsampled and multi-scale features are extracted: first, the X-band observation data is downsampled to reduce the spatial resolution to the same level as the S-band observation data, and the downsampled high-resolution input data HR Down Input is generated to simulate the data distribution under low-resolution input conditions; secondly, the high-resolution input data HR Down Input is input into the same hole multi-scale convolutional network structure as the low-resolution branch, and flows through the small receptive field convolution layer, the medium receptive field convolution layer and the large receptive field convolution layer in sequence to extract multi-scale features and generate a downsampled high-resolution HR Down feature map. Through the symmetrical feature extraction structure, the model learns the common features of low-resolution and high-resolution data.

[0026] Step 3.2: Build a data fusion module. The specific structure of the module is as follows: Figure 5 As shown: The data fusion module deeply fuses the features extracted from the two branches to generate a comprehensive feature map: first, a channel splicing operation is performed to splice the low-resolution LR feature map after multi-scale feature extraction with the downsampled high-resolution HR Down feature map according to the channel dimension. The spliced ​​feature map passes through the convolution layer, batch normalization layer and rectified linear unit in turn to achieve deep fusion of the features from the low-resolution LR and downsampled high-resolution HR Down data; The convolution layer in the data fusion module is a two-dimensional convolution layer, whose input is the spliced ​​128-channel feature map, and the convolution operation is performed by sliding a 3×3 convolution kernel on the feature map, and finally outputs a 64-channel feature map; the batch normalization layer is a two-dimensional batch normalization layer, which is used to normalize the data to a distribution with a mean of 0 and a variance of 1; the rectified linear unit is a ReLU activation function, which is used to perform nonlinear transformation on the batch-normalized feature map; the ReLU activation function is: , in is the input value after being processed by the batch normalization layer, is the output value after the ReLU activation function. If Greater than 0, output equal ;if Less than or equal to 0, output is equal to 0;

[0027] Step 3.3, construct an upsampling module: the upsampling module includes a channel attention module, two residual dense blocks, a convolutional layer and bicubic interpolation upsampling; The channel attention module, two residual dense blocks, a convolutional layer for expanding the number of channels, bicubic interpolation upsampling, and a convolutional layer for adjusting the number of channels are connected in sequence: first, the channel attention module can adaptively adjust the feature response of each channel to enhance the feature expression capability; followed by two residual dense blocks, which combine the advantages of residual and dense connection to alleviate the gradient vanishing problem and promote feature propagation reuse; then, through the convolutional layer for expanding the number of channels, the number of channels of the output feature map is expanded to provide sufficient information for subsequent upsampling operations; then, the bicubic interpolation function F.interpolate is used for upsampling to complete the conversion from low-resolution input to high-resolution output; finally, the convolutional layer for adjusting the number of channels is used to output the final upsampling result with the same number of channels as the input image; Among them, in the channel attention module, the spliced ​​feature map output by the data fusion module is first received; then the spliced ​​feature map is adaptively averaged and adaptively max-pooled respectively; then the results of average pooling and max-pooling are added together and input into a fully connected network consisting of two linear layers, ReLU activation function and Sigmoid function; then the shape of the attention weight is adjusted to the same size as the spliced ​​feature map; finally, the spliced ​​feature map is element-wise multiplied with the attention weight to obtain the feature map adjusted by the channel attention mechanism; The residual dense block performs the following operations: Step a1, during forward propagation, receive the feature map f adjusted by the channel attention mechanism; first, perform a convolution operation on the feature map f, and then activate it through the LeakyReLU function to obtain the feature map f1; then, concatenate the feature map f and the feature map f1 in the channel dimension, perform convolution and activate it through the LeakyReLU function to obtain the feature map f2; Step a2: According to step a1, the obtained feature maps f1 and f2 are continuously concatenated in the channel dimension and input into the convolution layer in the residual dense block to obtain feature maps f3 and f4 in turn; finally, all feature maps f1, f2, f3 and f4 are concatenated in the channel dimension and convolved to obtain feature map f5, and feature map f5 is multiplied by 0.2 and added to feature map f to obtain the output of the residual dense block; The formula for upsampling the bicubic interpolation function F.interpolate is: , in The function represents the use of the bicubic interpolation method to fused feature maps processed by the convolution layer with an expanded number of channels. Upsample to the specified size , and obtain a high-resolution feature map , complete the conversion from low-resolution input to high-resolution output; Indicates the height of the high-resolution image, Indicates the width of the high-resolution image; Step 3.4, modify the loss function: Generally, as the altitude increases, the quality and accuracy of the data received by the weather radar will be affected to a certain extent, and the data reliability will decrease accordingly. Therefore, the height weight adjustment loss is introduced to perform linear mapping based on the input altitude information, and associate the altitude with the weight. The lower the altitude, the higher the weight. At the same time, the Sobel gradient loss is introduced to enhance the edge information, including: In the loss calculation phase, the original L1 loss is calculated first, which reflects the difference between the model output and the true label; then the Sobel gradient loss is calculated, which measures the deviation between the Sobel gradient of the model output image and the Sobel gradient of the true label image; then these two losses are combined to get the total loss; then the total loss is multiplied by the height weight, and the results are averaged to finally get the weighted loss, so that samples with lower heights correspond to higher weights, which will have a greater impact in the loss calculation process. The weighted loss function formula is: , , in, represents the total loss, represents L1 loss, represents the Sobel gradient loss, represents the weight of Sobel gradient loss, n represents the number of samples, represents the total loss of the i-th sample, represents the height loss corresponding to the i-th sample, where i=1,2,…,n, represents the weighted loss.

[0028] By sequentially constructing the feature extraction module, data fusion module, and upsampling module, we finally obtained the spatial downscaling model.

[0029] Step 4: Train, verify and test the spatial downscaling model based on the preprocessed data. The idea is as follows: Figure 6 As shown, specifically including: Step 4.1, divide the data set: The S-band and X-band observation data after data preprocessing are taken as data pairs and divided into training set, validation set and test set in proportion (for example, 7:2:1); the S-band observation data are taken as low-resolution data, and the X-band observation data are taken as high-resolution radar data, and the low-resolution data and the high-resolution data correspond one to one.

[0030] Step 4.2: Train the spatial downscaling model based on the divided data set: In the training phase, m S-band observation data in the training set are input. and m X-band observation data , the output prediction result is the image after model space downscaling, represents the mth S-band observation data, Represents the mth X-band observation data, where m is the number of data in the training set, ranging from 7000 to 10000, covering data in different time periods.

[0031] According to the weighted loss function of the spatial downscaling model, the iterative weight parameters are updated through the back propagation algorithm, and this step is trained iteratively until the model converges.

[0032] Step 5, perform spatial downscaling of the S-band observation data according to the trained spatial downscaling model; The S-band observation data is spatially downscaled according to the trained spatial downscaling model; the data generated by the model is visualized and compared with the actual X-band observation data.

[0033] Step 6, determine whether the generated data is close enough to the true value; The comparison is based on the evaluation index structural similarity (SSIM), which is based on the brightness of the image. , Contrast and structure The calculation is done in three parts. For two images and , the SSIM formula is: , Among them, the brightness comparison function , contrast comparison function and structure comparison functions The calculation formulas are: , , , in represents X-band observation data, represents the S-band observation data, and They are The mean and The mean of and They are The standard deviation and The standard deviation of yes and The covariance of , and is a constant, is the weight coefficient, when , it means that when calculating SSIM, the brightness comparison function , contrast comparison function and structure comparison functions The weight of the impact on the final result is the same. When the two images are the same, .

[0034] Once the model is trained, it is used to spatially downscale the S-band radar data, and the data generated by the model is compared to the X-band radar data. If the generated data is close enough, the model can be put into use. Figure 7 This is an effect diagram of the multi-band radar data spatial downscaling method based on data fusion of the present invention. Figure 7 As shown in the figure, the S-band radar data is input, and after being processed by the model, high-resolution S-band radar data after spatial downscaling is obtained.

[0035] The present invention provides a multi-band radar data spatial downscaling method based on data fusion. There are many methods and ways to implement the technical solution. The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, without departing from the principle of the present invention, several improvements and modifications can be made, and these improvements and modifications should also be regarded as the protection scope of the present invention. All components not specified in this embodiment can be implemented by existing technologies.

Claims

1. A multi-band radar data spatial downscaling method based on data fusion, characterized in that: The following steps are involved: Step 1, obtaining S-band and X-band observation data from weather radar; Step 2, preprocessing the acquired observation data; Step 3, construct a spatial downscaling model; Step 4, training, verifying and testing the spatial downscaling model through the observation data after data preprocessing; Step 5, perform spatial downscaling of the S-band observation data according to the trained spatial downscaling model; Step 6: Determine whether the generated data is close to the true value.

2. The method according to claim 1, characterized in that Step 2 includes: Step 2.1, time-space matching of S-band and X-band observation data: According to the time identification information contained in the original observation file name, the X-band observation data and the S-band radar are time-calibrated so that the X-band observation data and the S-band radar are aligned in the time dimension; According to the latitude and longitude information corresponding to different stations, the S-band observation data is clipped according to the latitude and longitude values; Step 2.2, rotate and align the X-band observation data: By analyzing the historical observation data of the X-band, the current angle deviation value of the X-band is determined; Based on the angle deviation value, the radar scanning center is used as the rotation reference point, and a rotation matrix algorithm based on the trigonometric function principle is used to transform the coordinate system of the X-band observation data; Step 2.3, interpolate the X-band observation data to fill in the missing data: Based on the distribution characteristics of the valid data area acquired by the X-band radar, the data range of invalid values ​​in the data is calculated to define the scanning blind area range and boundary conditions where data is missing; Select an interpolation algorithm, use the valid data points around the scanning blind area as the reference data source, perform data estimation according to the established interpolation rules, and fill the estimated data into the corresponding missing positions; Step 2.4, remove low-quality data: perform quality assessment on the X-band observation data and the S-band observation data, and delete the low-quality radar data from the X-band observation data and the S-band observation data; use the X-band observation data as high-resolution input data and the S-band observation data as low-resolution input data.

3. The method according to claim 2, characterized in that Step 3 includes: Step 3.1, construct two independent feature extraction modules to process low-resolution input data and downsampled high-resolution input data respectively; Multi-scale feature extraction for low-resolution input data: The S-band observation data is input into the dilated multi-scale convolutional network, and flows through the small receptive field convolution layer, the medium receptive field convolution layer and the large receptive field convolution layer in sequence to extract multi-scale features and generate a low-resolution LR feature map; Down-sampling and multi-scale feature extraction of high-resolution input data: First, down-sampling operation is performed on X-band observation data to reduce the spatial resolution to the same level as S-band observation data, and down-sampled high-resolution input data HR Down Input is generated to simulate the data distribution under low-resolution input conditions; secondly, the high-resolution input data HRDown Input is input into the same hole multi-scale convolutional network structure as the low-resolution branch, and flows through the small receptive field convolution layer, the medium receptive field convolution layer and the large receptive field convolution layer in sequence to extract multi-scale features and generate a down-sampled high-resolution HRDown feature map. Through the symmetrical feature extraction structure, the model learns the common features of low-resolution and high-resolution data; Step 3.2, build data fusion module: The data fusion module deeply fuses the features extracted from the two branches to generate a comprehensive feature map: first, a channel splicing operation is performed to splice the low-resolution LR feature map after multi-scale feature extraction with the downsampled high-resolution HRDown feature map according to the channel dimension. The spliced ​​feature map passes through the convolution layer, batch normalization layer and rectified linear unit in turn to achieve deep fusion of the features from the low-resolution LR and downsampled high-resolution HR Down data; The convolution layer in the data fusion module is a two-dimensional convolution layer, whose input is the spliced ​​128-channel feature map, and the convolution operation is performed by sliding a 3×3 convolution kernel on the feature map, and finally outputs a 64-channel feature map; the batch normalization layer is a two-dimensional batch normalization layer, which is used to normalize the data to a distribution with a mean of 0 and a variance of 1; the rectified linear unit is a ReLU activation function, which is used to perform nonlinear transformation on the batch-normalized feature map; the ReLU activation function is: ， in is the input value after being processed by the batch normalization layer, is the output value after the ReLU activation function. If Greater than 0, output equal ;if Less than or equal to 0, output is equal to 0; Step 3.3, construct an upsampling module: the upsampling module includes a channel attention module, two residual dense blocks, a convolutional layer and bicubic interpolation upsampling; The channel attention module, two residual dense blocks, a convolutional layer for expanding the number of channels, bicubic interpolation upsampling, and a convolutional layer for adjusting the number of channels are connected in sequence: first, the channel attention module, followed by two residual dense blocks, and then the convolutional layer for expanding the number of channels, and then the bicubic interpolation function F.interpolate is used for upsampling to complete the conversion from low-resolution input to high-resolution output; finally, the convolutional layer for adjusting the number of channels is used to output the final upsampling result with the same number of channels as the input image; Among them, in the channel attention module, the spliced ​​feature map output by the data fusion module is first received; then the spliced ​​feature map is adaptively averaged and adaptively max-pooled respectively; then the results of average pooling and max-pooling are added together and input into a fully connected network consisting of two linear layers, ReLU activation function and Sigmoid function; then the shape of the attention weight is adjusted to the same size as the spliced ​​feature map; finally, the spliced ​​feature map is element-wise multiplied with the attention weight to obtain the feature map adjusted by the channel attention mechanism; The residual dense block performs the following operations: Step a1, during forward propagation, receive the feature map f adjusted by the channel attention mechanism; first, perform a convolution operation on the feature map f, and then activate it through the LeakyReLU function to obtain the feature map f1; then, concatenate the feature map f and the feature map f1 in the channel dimension, perform convolution and activate it through the LeakyReLU function to obtain the feature map f2; Step a2: According to step a1, the obtained feature maps f1 and f2 are continuously concatenated in the channel dimension and input into the convolution layer in the residual dense block to obtain feature maps f3 and f4 in turn; finally, all feature maps f1, f2, f3 and f4 are concatenated in the channel dimension and convolved to obtain feature map f5, and feature map f5 is multiplied by 0.2 and added to feature map f to obtain the output of the residual dense block; The formula for upsampling the bicubic interpolation function F.interpolate is: ， in The function represents the use of the bicubic interpolation method to fused feature maps processed by the convolution layer with an expanded number of channels. Upsample to the specified size , and obtain a high-resolution feature map , complete the conversion from low-resolution input to high-resolution output; Indicates the height of the high-resolution image, Indicates the width of the high-resolution image; Step 3.4, modify the loss function: introduce height weight adjustment loss, perform linear mapping based on the input height information, associate height with weight, and introduce Sobel gradient loss to enhance edge information, including: The weight loss function formula is: ， ， in, represents the total loss, represents L1 loss, represents the Sobel gradient loss, represents the weight of Sobel gradient loss, n represents the number of samples, represents the total loss of the i-th sample, represents the height loss corresponding to the i-th sample, where i=1,2,…,n, represents the weighted loss; By sequentially constructing the feature extraction module, data fusion module, and upsampling module, we finally obtained the spatial downscaling model.

4. The method according to claim 3, characterized in that Step 4 includes: Step 4.1, divide the data set: The S-band and X-band observation data after data preprocessing are used as data pairs and divided into training set, validation set and test set according to proportion; the S-band observation data are used as low-resolution data, and the X-band observation data are used as high-resolution radar data, and the low-resolution data and high-resolution data correspond one to one; Step 4.2: Train the spatial downscaling model based on the divided data set: In the training phase, m S-band observation data in the training set are input. and m X-band observation data , the output prediction result is the image after model space downscaling, represents the mth S-band observation data, Represents the mth X-band observation data, where m is the number of data in the training set.

5. The method according to claim 4, characterized in that Step 4.2 also includes: updating the iterative weight parameters according to the weighted loss function of the spatial downscaling model through the back propagation algorithm, and performing cyclic iterative training until the model converges.

6. The method according to claim 5, characterized in that In step 5, the S-band observation data is spatially downscaled according to the trained spatial downscaling model; the data generated by the spatial downscaling model is visualized and compared with the actual X-band observation data.

7. The method according to claim 6, characterized in that In step 6, the evaluation index structural similarity SSIM is used to determine whether the generated data is close to the true value. The formula is: ， Among them, the brightness comparison function , contrast comparison function and structure comparison functions The calculation formulas are: ， ， ， in represents X-band observation data, represents the S-band observation data, and They are The mean and The mean of and They are The standard deviation and The standard deviation of yes and The covariance of , and is a constant, is the weight coefficient.

8. The method according to claim 7, characterized in that In step 6, when , it means that when calculating SSIM, the brightness comparison function , contrast comparison function and structure comparison functions The weight of the impact on the final result is the same. When the two images are the same, .

9. An electronic device, characterized in that: The method comprises a processor and a memory, wherein the memory stores program codes, and when the program codes are executed by the processor, the processor executes the steps of the method according to any one of claims 1 to 8.

10. A storage medium, characterized in that: A computer program or instruction is stored, and when the computer program or instruction is run on a computer, the steps of the method according to any one of claims 1 to 8 are executed.

Images