Radar echo attenuation correction methods, devices, electronic equipment, media and products

Abstract

This invention provides a method, apparatus, electronic device, medium, and product for radar echo attenuation correction, relating to the field of radar technology. The method includes: acquiring the differential propagation phase shift rate of the radar; inputting the differential propagation phase shift rate into an echo attenuation correction model to obtain the echo attenuation correction result output by the model; wherein the echo attenuation correction model is trained based on differential propagation phase shift rate samples and attenuation rate samples, and is a model for performing radar echo attenuation correction. The technical solution of this invention processes the differential propagation phase shift rate of the radar using a trained echo attenuation correction model to obtain the radar echo attenuation correction result. The echo attenuation correction model has better stability and generalization ability, thereby improving the correction effect of the radar echo attenuation correction result.

Description

Technical Field

[0001] This invention relates to the field of radar technology, and in particular to a radar echo attenuation correction method, apparatus, electronic device, medium, and product. Background Technology

[0002] X-band radar, with its higher spatiotemporal resolution, shorter wavelength, and higher phase sensitivity, offers superior detection capabilities compared to conventional weather radar. Compared to ordinary Doppler weather radar, dual-polarization radar exhibits significant advantages in short-term nowcasting, precipitation estimation, and phase inversion. However, compared to C-band and S-band radar, X-band radar operates at a higher frequency, resulting in more pronounced atmospheric attenuation. Therefore, the most crucial aspect of effectively utilizing X-band radar lies in attenuation correction.

[0003] Currently, in existing technologies, traditional attenuation correction is mainly based on empirical formulas. By collecting fixed empirical coefficients and parameters and substituting them into the empirical formulas, the attenuation correction result is obtained.

[0004] Therefore, traditional empirical formulas for attenuation correction of radar echo signals have limited parameters, poor generalization ability, and great uncertainty due to the use of fixed empirical coefficients, resulting in inaccurate correction results. Summary of the Invention

[0005] This invention provides a radar echo attenuation correction method, apparatus, electronic device, medium, and product to address the shortcomings of existing technologies that use traditional empirical formulas to correct radar echo signal attenuation. These formulas suffer from limited parameters, poor generalization ability, and significant uncertainty due to the use of fixed empirical coefficients, resulting in inaccurate correction results. The invention achieves this by processing the differential propagation phase shift rate of the radar using a trained echo attenuation correction model, thereby obtaining the corrected radar echo attenuation. This model exhibits better stability and generalization ability, thus improving the correction effect of the radar echo attenuation correction results.

[0006] This invention provides a radar echo attenuation correction method, comprising the following steps.

[0007] Obtain the differential propagation phase shift rate of the radar.

[0008] Input the differential propagation phase shift rate into the echo attenuation correction model to obtain the echo attenuation correction result output by the echo attenuation correction model; wherein, the echo attenuation correction model is trained based on differential propagation phase shift rate samples and attenuation rate samples, and the echo attenuation correction model is a model for performing echo attenuation correction of radar.

[0009] According to the radar echo attenuation correction method provided by the present invention, the echo attenuation correction model is trained based on the following steps: obtaining differential propagation phase shift rate samples and attenuation rate samples; determining the network architecture based on the differential propagation phase shift rate samples; and optimizing and training the network architecture based on the attenuation rate samples to obtain the echo attenuation correction model.

[0010] According to a radar echo attenuation correction method provided by the present invention, obtaining an attenuation rate sample includes: obtaining an initial first reflectivity factor set for a radar in a first band and an initial second reflectivity factor set for a radar in a second band; wherein the first band and the second band are different bands; determining a first reflectivity factor set and a second reflectivity factor set based on the initial first reflectivity factor set and the initial second reflectivity factor set; wherein the first reflectivity factor set is a factor set after preprocessing the initial first reflectivity factor set, and the second reflectivity factor set is a factor set after preprocessing the initial second reflectivity factor set; interpolating the second reflectivity factor set into the first reflectivity factor set to obtain an attenuation rate sample.

[0011] According to a radar echo attenuation correction method provided by the present invention, determining a first reflectivity factor set and a second reflectivity factor set based on an initial first reflectivity factor set and an initial second reflectivity factor set includes: obtaining all first correlation coefficients corresponding to all initial first reflectivity factors in the initial first reflectivity factor set, and determining all second correlation coefficients corresponding to initial second reflectivity factors in the initial second reflectivity factor set; determining whether all first correlation coefficients and all second correlation coefficients are less than preset coefficients; removing the initial first reflectivity factors corresponding to the first correlation coefficients less than the preset coefficients from the initial first reflectivity factor set to obtain a candidate first reflectivity factor set, and removing the initial second reflectivity factors corresponding to the second correlation coefficients less than the preset coefficients from the initial second reflectivity factor set to obtain a candidate second reflectivity factor set; performing data unification processing on all initial first reflectivity factors in the candidate first reflectivity factor set to obtain a first reflectivity factor set, and performing data unification processing on all initial second reflectivity factors in the candidate second reflectivity factor set to obtain a second reflectivity factor set.

[0012] According to the radar echo attenuation correction method provided by the present invention, the network architecture is determined based on differential propagation phase shift rate samples, including: inputting differential propagation phase shift rate samples into the encoder of the infrastructure to obtain the encoding result output by the encoder of the infrastructure; wherein, the infrastructure is a pre-defined basic network architecture; inputting the encoding result into the decoder of the infrastructure to obtain the decoding result output by the decoder of the infrastructure; inputting the decoding result into the fully connected layer of the infrastructure to obtain the attenuation result output by the fully connected layer of the infrastructure; and further optimizing and training the infrastructure based on the attenuation result to determine the network architecture.

[0013] According to the radar echo attenuation correction method provided by the present invention, the differential propagation phase shift rate sample is input into the encoder of the infrastructure to obtain the encoding result output by the encoder of the infrastructure, including: inputting the differential propagation phase shift rate sample into the linear coding layer in the encoder to obtain the coding features output by the linear coding layer; and inputting the coding features into the position coding layer in the encoder to obtain the encoding result output by the position coding layer.

[0014] The present invention also provides a radar echo attenuation correction device, comprising the following modules.

[0015] The acquisition module is used to acquire the differential propagation phase shift rate of the radar.

[0016] The correction module is used to input the differential propagation phase shift rate into the echo attenuation correction model and obtain the echo attenuation correction result output by the echo attenuation correction model. The echo attenuation correction model is trained based on differential propagation phase shift rate samples and attenuation rate samples. The echo attenuation correction model is a model for performing echo attenuation correction on radar.

[0017] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the echo attenuation correction method of any of the radars described above.

[0018] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the echo attenuation correction method for any of the radars described above.

[0019] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the echo attenuation correction method for any of the radars described above.

[0020] This invention provides a radar echo attenuation correction method, apparatus, electronic device, medium, and product. It obtains the differential propagation phase shift rate of the radar; inputs the differential propagation phase shift rate into an echo attenuation correction model to obtain the echo attenuation correction result output by the model; wherein, the echo attenuation correction model is trained based on differential propagation phase shift rate samples and attenuation rate samples, and is a model for radar echo attenuation correction. The technical solution of this invention addresses the shortcomings of existing technologies that use traditional empirical formulas for radar echo signal attenuation correction. Due to the use of fixed empirical coefficients, the parameters are limited, the generalization ability is poor, and there is significant uncertainty, leading to inaccurate correction results. This invention achieves radar echo attenuation correction results by processing the differential propagation phase shift rate of the radar using a trained echo attenuation correction model. The echo attenuation correction model has better stability and generalization ability, thereby improving the correction effect of radar echo attenuation. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0022] Figure 1 This is a flowchart illustrating the radar echo attenuation correction method provided by the present invention.

[0023] Figure 2 This is a schematic diagram of the radar echo attenuation correction device provided by the present invention.

[0024] Figure 3 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0026] The following is combined with Figure 1The radar echo attenuation correction method provided by this invention is described below. This radar echo attenuation correction method is applicable to the echo attenuation correction of X-band dual-polarization radar. The execution subject of this method can be an electronic device or a radar echo attenuation correction device installed in the electronic device. The radar echo attenuation correction device can be implemented by software, hardware or a combination of both. Figure 1 This is a flowchart illustrating the radar echo attenuation correction method provided by the present invention, as shown below. Figure 1 As shown, the method includes the following steps 101 and 102.

[0027] Step 101: Obtain the differential propagation phase shift rate of the radar.

[0028] In this step, the radar is an electronic device that uses electromagnetic waves to detect targets. The radar can be, for example, an X-band dual-polarization radar; this embodiment does not limit this to a specific type.

[0029] The differential propagation phase shift rate is a parameter used in weather radar to describe the influence of precipitation particles on the propagation of electromagnetic waves. The differential propagation phase shift rate represents the rate of change of the phase difference between horizontal and vertical polarized waves on the propagation path per unit distance. For example, it can be the rate of change of the three elevation angle layers of radar at 1.5°, 2.4°, and 3.4°. This embodiment does not limit this.

[0030] Specifically, the differential propagation phase of the radar within a unit distance is obtained, and the differential propagation phase shift rate of the radar is obtained based on the ratio of the differential propagation phase of the radar to that within a unit distance.

[0031] Step 102: Input the differential propagation phase shift rate into the return attenuation correction model to obtain the return attenuation correction result output by the return attenuation correction model.

[0032] In this step, the echo attenuation correction model is trained based on the differential propagation phase shift rate sample and attenuation rate sample. The echo attenuation correction model is the model for correcting the echo attenuation of the radar.

[0033] The attenuation rate samples can be, for example, horizontal attenuation rate samples, differential attenuation rate samples, etc., and the return attenuation correction results can be, for example, horizontal reflectivity factor attenuation correction results, differential reflectivity factor attenuation correction results; this embodiment does not limit these. The model can be trained based on differential propagation phase shift rate samples and horizontal attenuation rate samples so that the resulting return attenuation correction model can analyze the input differential propagation phase shift rate, thereby obtaining the horizontal reflectivity factor attenuation correction result; alternatively, the model can be trained based on differential propagation phase shift rate samples and differential attenuation rate samples so that the resulting return attenuation correction model can analyze the input differential propagation phase shift rate, thereby obtaining the differential reflectivity factor attenuation correction result.

[0034] Specifically, after obtaining the trained echo attenuation correction model, the differential propagation phase shift rate is used as the input of the echo attenuation correction model. Then, the echo attenuation correction model is trained on the input differential propagation phase shift rate to obtain the echo attenuation correction result output by the echo attenuation correction model.

[0035] For example, when the attenuation rate sample is determined to be a horizontal attenuation rate sample, the model can be trained using the differential propagation phase shift rate sample and the horizontal attenuation rate sample as inputs to the return attenuation correction model. This allows the obtained return attenuation correction model to process the differential propagation phase shift rate, resulting in the horizontal reflectivity factor attenuation correction result output by the return attenuation correction model. When the attenuation rate sample is determined to be a differential attenuation rate sample, the model can be trained using the differential propagation phase shift rate sample and the differential attenuation rate sample as inputs to the return attenuation correction model. This allows the obtained return attenuation correction model to process the differential propagation phase shift rate, resulting in the differential reflectivity factor attenuation correction result output by the return attenuation correction model. Training the model with different attenuation rate samples, such as horizontal and differential attenuation rate samples, will result in different output results obtained by the final return attenuation correction model after processing the differential propagation phase shift rate. The output results are related to different attenuation rate samples, such as horizontal and differential attenuation rate samples.

[0036] In one specific implementation, the echo attenuation correction model is trained based on the following steps: obtaining differential propagation phase shift rate samples and attenuation rate samples; determining the network architecture based on the differential propagation phase shift rate samples; and optimizing and training the network architecture based on the attenuation rate samples to obtain the echo attenuation correction model.

[0037] In this step, the differential propagation phase shift rate samples are used for network construction. The attenuation rate samples are used for optimizing the constructed network architecture during training.

[0038] Specifically, the differential propagation phase shift rate samples are obtained, the network architecture is determined based on the differential propagation phase shift rate samples, and the attenuation rate samples are obtained. The network architecture is then optimized and trained based on the attenuation rate samples to obtain the echo attenuation correction model.

[0039] In one specific embodiment, obtaining the attenuation rate sample includes: obtaining an initial first reflectivity factor set for a radar in a first band and an initial second reflectivity factor set for a radar in a second band; wherein the first band and the second band are different bands; determining a first reflectivity factor set and a second reflectivity factor set based on the initial first reflectivity factor set and the initial second reflectivity factor set; wherein the first reflectivity factor set is a factor set after preprocessing the initial first reflectivity factor set, and the second reflectivity factor set is a factor set after preprocessing the initial second reflectivity factor set; interpolating the second reflectivity factor set into the first reflectivity factor set to obtain the attenuation rate sample.

[0040] In this step, the first band may be, for example, the X-band of a dual-polarization radar, and the second band may be, for example, the S-band of a new generation weather radar. This embodiment does not limit this.

[0041] The initial first reflectivity factor set is the set of first reflectivity factors at different distances in the X-band of the dual-polarization radar at a range of 150 km. The initial first reflectivity factor set may be, for example, the initial first horizontal reflectivity factor set or the initial first differential reflectivity factor set. This embodiment does not limit this.

[0042] The range library refers to the small units divided along the ray direction in radar echo signal processing according to distance. The number of range libraries can be, for example, 15, but this embodiment does not limit this.

[0043] The initial second reflectivity factor set is a set of second reflectivity factors on different ranges of the S-band of the new generation weather radar. The initial second reflectivity factor set may be, for example, an initial second horizontal reflectivity factor set or an initial second differential reflectivity factor set. This embodiment does not limit this.

[0044] Preprocessing may include quality control and normalization of the initial first reflectivity factor set and the initial second reflectivity factor set, respectively. Quality control may include filtering out non-meteorological echoes, ground features, and isolated clutter, but this embodiment does not limit this.

[0045] The attenuation rate sample is the true echo value. The attenuation rate sample can be, for example, a horizontal attenuation rate sample or a differential attenuation rate sample. Horizontal attenuation rate sample The calculation is shown in formula (1).

[0046] (1)

[0047] In formula (1), This represents the distance library number, for example, it could be 1, 2, 3… This indicates the distance at which the data was collected, for example, 1.125 kilometers. The second horizontal reflectivity factor represents a range library of the S-band of a new generation of weather radar. The first horizontal reflectivity factor represents a range library in the X-band of a dual-polarization radar.

[0048] Differential decay rate sample The calculation is shown in formula (2).

[0049] (2)

[0050] In formula (2), This represents the distance library number, for example, it could be 1, 2, 3… This indicates the distance at which the data was collected, for example, 1.125 kilometers. The second differential reflectivity factor represents a range library of a new generation of weather radar in the S-band. The first differential reflectivity factor represents a certain range library in the X-band of the dual-polarization radar; this embodiment does not limit this factor.

[0051] In one specific embodiment, obtaining the attenuation rate sample includes: obtaining an initial first reflectivity factor set for a radar in a first band and an initial second reflectivity factor set for a radar in a second band; wherein the first band and the second band are different bands; determining a first reflectivity factor set and a second reflectivity factor set based on the initial first reflectivity factor set and the initial second reflectivity factor set; wherein the first reflectivity factor set is a factor set after preprocessing the initial first reflectivity factor set, and the second reflectivity factor set is a factor set after preprocessing the initial second reflectivity factor set; interpolating the second reflectivity factor set into the first reflectivity factor set to obtain the attenuation rate sample.

[0052] Specifically, in the first band, all initial first reflectivity factors of the radar in different range databases are acquired, and a first reflectivity factor set is determined based on all first reflectivity factors. In the second band, all initial second reflectivity factors of the radar in different range databases are acquired, and a second reflectivity factor set is determined based on all second reflectivity factors. Then, the initial first reflectivity factor set and the initial second reflectivity factor set are preprocessed, including quality control and normalization, to obtain a preprocessed first reflectivity factor set and a preprocessed second reflectivity factor set. Finally, the preprocessed second reflectivity factor set is interpolated into the preprocessed first reflectivity factor set using bilinear interpolation to obtain the attenuation rate sample.

[0053] In one specific embodiment, determining the first reflectance factor set and the second reflectance factor set based on the initial first reflectance factor set and the initial second reflectance factor set includes: obtaining all first correlation coefficients corresponding to all initial first reflectance factors in the initial first reflectance factor set, and determining all second correlation coefficients corresponding to the initial second reflectance factors in the initial second reflectance factor set; determining whether all first correlation coefficients and all second correlation coefficients are less than preset coefficients; removing the initial first reflectance factors corresponding to the first correlation coefficients less than the preset coefficients from the initial first reflectance factor set to obtain a candidate first reflectance factor set, and removing the initial second reflectance factors corresponding to the second correlation coefficients less than the preset coefficients from the initial second reflectance factor set to obtain a candidate second reflectance factor set; performing data unification processing on all initial first reflectance factors in the candidate first reflectance factor set to obtain a first reflectance factor set, and performing data unification processing on all initial second reflectance factors in the candidate second reflectance factor set to obtain a second reflectance factor set.

[0054] In this step, both the first and second correlation coefficients represent the correlation between horizontally polarized and vertically polarized echoes within the radar resolution volume, ranging from [0, 1]. A value closer to 1 indicates better quality for all initial first reflectivity factor data in the initial first reflectivity factor set and all initial second reflectivity factor data in the initial second reflectivity factor set. The first and second correlation coefficients can be used to determine the data quality of all initial first reflectivity factor data in the initial first reflectivity factor set and all initial second reflectivity factor data in the initial second reflectivity factor set. Low first and second correlation coefficients may indicate the presence of non-meteorological echoes or noise in all initial first and second reflectivity factor data in the initial first and second reflectivity factor sets, suggesting non-uniform phase or noise in the corresponding range reservoir area.

[0055] The first correlation coefficient can be, for example, 0.95, and the second correlation coefficient can be, for example, 0.95. This embodiment does not limit this.

[0056] The preset coefficient is a pre-set coefficient used to judge the first correlation coefficient and the second correlation coefficient. For example, it can be 0.9. This embodiment does not limit this.

[0057] Specifically, the process involves obtaining all first correlation coefficients corresponding to all initial first reflectivity factors in the initial first reflectivity factor set, and determining all second correlation coefficients corresponding to the initial second reflectivity factors in the initial second reflectivity factor set; determining whether all first correlation coefficients and all second correlation coefficients are less than preset coefficients; removing the initial first reflectivity factors corresponding to the first correlation coefficients less than the preset coefficients from the initial first reflectivity factor set to obtain a candidate first reflectivity factor set, and removing the initial second reflectivity factors corresponding to the second correlation coefficients less than the preset coefficients from the initial second reflectivity factor set to obtain a candidate second reflectivity factor set; performing data unification processing on all initial first reflectivity factors in the candidate first reflectivity factor set to obtain a first reflectivity factor set, and performing data unification processing on all initial second reflectivity factors in the candidate second reflectivity factor set to obtain a second reflectivity factor set.

[0058] For example, the data unification process for all initial first reflectivity factors in the candidate first reflectivity factor set can be normalized according to the maximum and minimum values ​​of the -20 to 70 dB reflectivity factor (dBZ); the data unification process for all initial second reflectivity factors in the candidate second reflectivity factor set can be normalized according to the maximum and minimum values ​​of the -2 to 6 dB reflectivity factor.

[0059] In one specific implementation, determining the network architecture based on the differential propagation phase shift rate samples includes: inputting the differential propagation phase shift rate samples into the encoder of the infrastructure to obtain the encoding result output by the encoder of the infrastructure; wherein, the infrastructure is a pre-defined basic network architecture; inputting the encoding result into the decoder of the infrastructure to obtain the decoding result output by the decoder of the infrastructure; inputting the decoding result into the fully connected layer of the infrastructure to obtain the attenuation result output by the fully connected layer of the infrastructure; and further optimizing and training the infrastructure based on the attenuation result to determine the network architecture.

[0060] In this step, before determining the network architecture, the differential propagation phase shift rate samples can be uniformly configured to improve the accuracy of subsequent model training. The basic architecture is a pre-defined, untrained basic network architecture, which includes an encoder, decoder, and fully connected layers; this embodiment does not limit this.

[0061] The infrastructure includes an encoder, a decoder, and finally outputs the echo attenuation correction result through a fully connected layer. The infrastructure can be, for example, an X-band radar echo attenuation correction framework designed based on Transformer (a model architecture), but this embodiment does not limit it.

[0062] The decoder incorporates a masking mechanism to improve the model's robustness. Then, through parallel computation using a multi-head attention mechanism (e.g., 8 heads), the model can simultaneously focus on different parts of the input differential propagation phase shift rate sample sequence, learning information from multiple subspaces. Residual connections and normalization are added to improve convergence speed and promote better feature learning. The results are then passed to a feedforward neural network for further nonlinear mapping and feature extraction. Two decoding modules (2×) are set up for cyclic decoding. Finally, the decoding results are input into a fully connected layer to obtain the output attenuation result.

[0063] Specifically, the differential propagation phase shift rate samples are input into the encoder of the infrastructure to obtain the encoding result output by the encoder; the encoding result is input into the decoder of the infrastructure to obtain the decoding result output by the decoder; the decoding result is input into the fully connected layer of the infrastructure to obtain the attenuation result output by the fully connected layer; the infrastructure is further optimized and trained based on the attenuation result to determine the network architecture.

[0064] In one specific implementation, inputting the differential propagation phase shift rate sample into the encoder of the infrastructure to obtain the encoding result output by the encoder of the infrastructure includes: inputting the differential propagation phase shift rate sample into the linear coding layer in the encoder to obtain the encoding features output by the linear coding layer; and inputting the encoding features into the position coding layer in the encoder to obtain the encoding result output by the position coding layer.

[0065] In this step, the linear coding layer can be, for example, a layer that performs linear coding using 32-dimensional vectors and 64-dimensional vectors respectively; this embodiment does not limit this.

[0066] Specifically, the differential propagation phase shift rate sample is input into the linear coding layer in the encoder, and encoded through two linear coding layers to obtain the coding features output by the linear coding layer; the coding features are input into the position coding layer in the encoder, and the position coding layer supplements the position information to obtain the coding result output by the position coding layer.

[0067] For example, the effective values ​​of the difference propagation phase shift rate samples from 15 distance libraries can be obtained as input. These samples are encoded into a 64-dimensional vector through two linear layers in the encoder, and a positional encoding layer is used to supplement the positional information of the difference propagation phase shift rate sample data, enabling correlation between the preceding and following data. Then, a masking mechanism is introduced in the decoder to improve the model's robustness. Parallel computation through a multi-head attention mechanism allows the model to simultaneously focus on different parts of the difference propagation phase shift rate sample sequence, learning information from multiple subspaces. Residual connections and normalization are added to improve convergence speed and promote better feature learning. This data is then passed to a feedforward neural network for further nonlinear mapping and feature extraction. Two decoding modules (2×) are set up for cyclic decoding steps. Finally, the data is input into a fully connected layer to obtain the final attenuation result, and the echo attenuation correction result is obtained based on the attenuation result. The activation function used is the Rectified Linear Unit (ReLU), and Dropout is used to prevent overfitting. This embodiment does not limit this specific usage.

[0068] In one specific implementation, after obtaining the horizontal reflectivity factor attenuation correction result or the differential reflectivity factor attenuation correction result, the horizontal reflectivity factor attenuation correction result or the differential reflectivity factor attenuation correction result can be further evaluated to obtain the evaluation result.

[0069] Evaluation indicators such as ratio deviation, root mean square error, and mean absolute error are calculated based on the horizontal reflectivity factor attenuation correction results or the differential reflectivity factor attenuation correction results. The horizontal reflectivity factor attenuation correction results or the differential reflectivity factor attenuation correction results are evaluated based on the ratio deviation, root mean square error, and mean absolute error.

[0070] Among them, ratio deviation The calculation is shown in formula (3), root mean square error The calculation is shown in formula (4), the mean absolute error The calculation is shown in formula (5).

[0071] (3)

[0072] (4)

[0073] (5)

[0074] In formulas (3), (4), and (5), Represents the second set of reflectivity factors The second reflectivity factor corresponding to each distance library This indicates the corrected horizontal reflectivity factor attenuation result or the differential reflectivity factor attenuation result. This represents the total number of ranges in the range library. BIAS measures the difference between X and S segments; the closer its value is to 1, the smaller the difference in intensity between the large and small radars. RMSE measures the dispersion of intensity differences; the smaller its value, the smaller the difference in intensity between the large and small radars. MAE represents the average level of the absolute difference between echoes, with less influence on outliers; the smaller the MAE value, the higher the absolute accuracy of the horizontal reflectivity factor attenuation correction result or the differential reflectivity factor attenuation correction result.

[0075] For example, as shown in Table 1, the corrected horizontal reflectivity factor / differential reflectivity factor in the return attenuation correction result before correction and after correction by the return attenuation correction model are compared with the horizontal reflectivity factor / differential reflectivity factor corrected by the existing empirical formula. It is determined that the result of correction by the return attenuation correction model has improved the correction effect of the empirical formula and improved the correction accuracy.

[0076] Table 1

[0077]

[0078] As shown in Table 1, by comparing the BIAS before and after the horizontal reflectance factor correction, the model improved the BIAS from 0.875 to 0.972, which is better than the 0.901 after the empirical formula correction. For the differential reflectance factor, the BIAS before correction was 0.862, after the model correction it was 1.141, while after the empirical formula correction it was 1.273, indicating an overcorrection problem. The RMSE and MAE were 8.693 and 6.292 respectively before the horizontal reflectance factor correction, after the model correction they were 5.811 and 4.222, respectively, an improvement of 33.15% and 32.89%, respectively. After the empirical formula correction they were 6.820 and 5.113, respectively, an improvement of 21.54% and 18.73%, respectively. Before correction using the differential reflectance factor, RMSE and MAE were 1.679 and 1.271, respectively. After model correction, they were 0.972 and 0.697, representing improvements of 42.10% and 45.16%, respectively. After correction using the empirical formula, the values ​​were 1.382 and 1.008, representing improvements of 17.69% and 20.69%, respectively. All evaluation indicators showed significant improvements compared to the empirical formula correction method.

[0079] This invention provides a radar echo attenuation correction method. It obtains the differential propagation phase shift rate of the radar; inputs the differential propagation phase shift rate into an echo attenuation correction model to obtain the echo attenuation correction result output by the model. The echo attenuation correction model is trained based on differential propagation phase shift rate samples and attenuation rate samples, and is used to correct radar echo attenuation. Based on the above embodiments, the technical solution of this invention addresses the shortcomings of traditional empirical formulas for attenuation correction of radar echo signals. These formulas use fixed empirical coefficients, have limited parameters, poor generalization ability, and significant uncertainty, leading to inaccurate correction results. This invention achieves radar echo attenuation correction by processing the differential propagation phase shift rate of the radar using a trained echo attenuation correction model. The echo attenuation correction model has better stability and generalization ability, thus improving the correction effect of radar echo attenuation.

[0080] The radar echo attenuation correction device provided by the present invention will be described below. The radar echo attenuation correction device described below can be referred to in correspondence with the radar echo attenuation correction method described above.

[0081] Figure 2 This is a schematic diagram of the radar echo attenuation correction device provided by the present invention, with reference to... Figure 2 As shown, the radar echo attenuation correction device 200 includes an acquisition module 301 and a correction module 302.

[0082] The acquisition module 301 is used to acquire the differential propagation phase shift rate of the radar.

[0083] The correction module 302 is used to input the differential propagation phase shift rate into the echo attenuation correction model to obtain the echo attenuation correction result output by the echo attenuation correction model; wherein, the echo attenuation correction model is trained based on differential propagation phase shift rate samples and attenuation rate samples, and the echo attenuation correction model is a model for performing echo attenuation correction of radar.

[0084] In one example embodiment, the device further includes a model training module. The model training module is configured to: acquire differential propagation phase shift rate samples and acquire attenuation rate samples; determine the network architecture based on the differential propagation phase shift rate samples; and optimize and train the network architecture based on the attenuation rate samples to obtain an echo attenuation correction model.

[0085] In one example embodiment, the model training module obtains attenuation rate samples, specifically by: obtaining an initial first reflectivity factor set for a radar in a first band and an initial second reflectivity factor set for a radar in a second band; wherein the first band and the second band are different bands; determining a first reflectivity factor set and a second reflectivity factor set based on the initial first reflectivity factor set and the initial second reflectivity factor set; wherein the first reflectivity factor set is a factor set after preprocessing the initial first reflectivity factor set, and the second reflectivity factor set is a factor set after preprocessing the initial second reflectivity factor set; interpolating the second reflectivity factor set into the first reflectivity factor set to obtain attenuation rate samples.

[0086] In one example embodiment, the model training module determines the first reflectance factor set and the second reflectance factor set based on the initial first reflectance factor set and the initial second reflectance factor set. Specifically, this involves: obtaining all first correlation coefficients corresponding to all initial first reflectance factors in the initial first reflectance factor set, and determining all second correlation coefficients corresponding to the initial second reflectance factors in the initial second reflectance factor set; determining whether all first correlation coefficients and all second correlation coefficients are less than a preset coefficient; removing the initial first reflectance factors corresponding to the first correlation coefficients less than the preset coefficient from the initial first reflectance factor set to obtain a candidate first reflectance factor set, and removing the initial second reflectance factors corresponding to the second correlation coefficients less than the preset coefficient from the initial second reflectance factor set to obtain a candidate second reflectance factor set; performing data unification processing on all initial first reflectance factors in the candidate first reflectance factor set to obtain a first reflectance factor set, and performing data unification processing on all initial second reflectance factors in the candidate second reflectance factor set to obtain a second reflectance factor set.

[0087] In one example embodiment, the model training module determines the network architecture based on the difference propagation phase shift rate samples. Specifically, it is used to: input the difference propagation phase shift rate samples into the encoder of the infrastructure to obtain the encoding result output by the encoder of the infrastructure; wherein, the infrastructure is a pre-defined basic network architecture; input the encoding result into the decoder of the infrastructure to obtain the decoding result output by the decoder of the infrastructure; input the decoding result into the fully connected layer of the infrastructure to obtain the attenuation result output by the fully connected layer of the infrastructure; and further optimize and train the infrastructure based on the attenuation result to determine the network architecture.

[0088] In one example embodiment, the model training module inputs the differential propagation phase shift rate samples into the encoder of the infrastructure to obtain the encoding result output by the encoder of the infrastructure. Specifically, it is used to: input the differential propagation phase shift rate samples into the linear coding layer in the encoder to obtain the encoding features output by the linear coding layer; and input the encoding features into the position coding layer in the encoder to obtain the encoding result output by the position coding layer.

[0089] The apparatus of this embodiment can be used to execute the method of any embodiment in the radar echo attenuation correction method side embodiment. Its specific implementation process and technical effects are similar to those in the radar echo attenuation correction method side embodiment. For details, please refer to the detailed description in the radar echo attenuation correction method side embodiment, which will not be repeated here.

[0090] Figure 3 This is a schematic diagram of the structure of the electronic device provided by the present invention, such as... Figure 3 As shown, the electronic device may include a processor 310, a communications interface 320, a memory 330, and a communication bus 340. The processor 310, communications interface 320, and memory 330 communicate with each other via the communication bus 340. The processor 310 can call logic instructions in the memory 330 to execute a radar echo attenuation correction method. This method includes: obtaining the differential propagation phase shift rate of the radar; inputting the differential propagation phase shift rate into an echo attenuation correction model to obtain the echo attenuation correction result output by the echo attenuation correction model; wherein the echo attenuation correction model is trained based on differential propagation phase shift rate samples and attenuation rate samples, and the echo attenuation correction model is a model for performing radar echo attenuation correction.

[0091] Furthermore, the logical instructions in the aforementioned memory 330 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0092] On the other hand, the present invention also provides a computer program product, which includes a computer program that 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 radar echo attenuation correction method provided by the above methods. The method includes: obtaining the differential propagation phase shift rate of the radar; inputting the differential propagation phase shift rate into the echo attenuation correction model to obtain the echo attenuation correction result output by the echo attenuation correction model; wherein the echo attenuation correction model is obtained by training based on differential propagation phase shift rate samples and attenuation rate samples, and the echo attenuation correction model is a model for performing radar echo attenuation correction.

[0093] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the radar echo attenuation correction method provided by the above methods. The method includes: obtaining the differential propagation phase shift rate of the radar; inputting the differential propagation phase shift rate into an echo attenuation correction model to obtain an echo attenuation correction result output by the echo attenuation correction model; wherein the echo attenuation correction model is obtained by training based on differential propagation phase shift rate samples and attenuation rate samples, and the echo attenuation correction model is a model for performing radar echo attenuation correction.

[0094] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0095] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0096] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A radar echo attenuation correction method, characterized in that, include: Obtain the differential propagation phase shift rate of the radar; The differential propagation phase shift rate is input into the echo attenuation correction model to obtain the echo attenuation correction result output by the echo attenuation correction model; wherein, the echo attenuation correction model is trained based on differential propagation phase shift rate samples and attenuation rate samples, and the echo attenuation correction model is a model for radar echo attenuation correction; the echo attenuation correction model is trained based on the following steps: acquiring the differential propagation phase shift rate samples and acquiring the attenuation rate samples; determining the network architecture based on the differential propagation phase shift rate samples; optimizing and training the network architecture based on the attenuation rate samples to obtain the echo attenuation correction model; the acquisition of the attenuation rate samples includes: acquiring the first wave The radar system generates an initial first reflectivity factor set for a certain band and an initial second reflectivity factor set for a certain second band; wherein the first band and the second band are different bands; a first reflectivity factor set and a second reflectivity factor set are determined based on the initial first reflectivity factor set and the initial second reflectivity factor set; wherein the first reflectivity factor set is a factor set after preprocessing the initial first reflectivity factor set, and the second reflectivity factor set is a factor set after preprocessing the initial second reflectivity factor set; the second reflectivity factor set is interpolated into the first reflectivity factor set to obtain the attenuation rate sample.

2. The radar echo attenuation correction method according to claim 1, characterized in that, The step of determining the first reflectivity factor set and the second reflectivity factor set based on the initial first reflectivity factor set and the initial second reflectivity factor set includes: Obtain all first correlation coefficients corresponding to all initial first reflectivity factors in the initial first reflectivity factor set, and determine all second correlation coefficients corresponding to all initial second reflectivity factors in the initial second reflectivity factor set; Determine whether all the first correlation coefficients and all the second correlation coefficients are less than a preset coefficient; The initial first reflectivity factor corresponding to the first correlation coefficient that is less than the preset coefficient is removed from the initial first reflectivity factor set to obtain a candidate first reflectivity factor set, and the initial second reflectivity factor corresponding to the second correlation coefficient that is less than the preset coefficient is removed from the initial second reflectivity factor set to obtain a candidate second reflectivity factor set. The first reflectance factor set is obtained by performing data unification processing on all the initial first reflectance factors in the candidate first reflectance factor set, and the second reflectance factor set is obtained by performing data unification processing on all the initial second reflectance factors in the candidate second reflectance factor set.

3. The radar echo attenuation correction method according to claim 1, characterized in that, The step of determining the network architecture based on the differential propagation phase shift rate samples includes: The differential propagation phase shift rate sample is input into the encoder of the infrastructure to obtain the encoding result output by the encoder of the infrastructure; wherein, the infrastructure is a pre-defined basic network architecture; The encoding result is input into the decoder of the infrastructure to obtain the decoding result output by the decoder of the infrastructure. The decoding result is input into the fully connected layer of the infrastructure to obtain the attenuation result output by the fully connected layer of the infrastructure; Based on the attenuation results, the infrastructure is further optimized and trained to determine the network architecture.

4. The radar echo attenuation correction method according to claim 3, characterized in that, The step of inputting the differential propagation phase shift rate sample into the encoder of the infrastructure to obtain the encoding result output by the encoder of the infrastructure includes: The differential propagation phase shift rate sample is input into the linear coding layer in the encoder to obtain the coding features output by the linear coding layer; The encoded features are input into the position encoding layer in the encoder to obtain the encoded result output by the position encoding layer.

5. A radar echo attenuation correction device, characterized in that, include: The acquisition module is used to acquire the differential propagation phase shift rate of the radar; A correction module is used to input the differential propagation phase shift rate into an echo attenuation correction model to obtain the echo attenuation correction result output by the echo attenuation correction model; wherein, the echo attenuation correction model is trained based on differential propagation phase shift rate samples and attenuation rate samples, and the echo attenuation correction model is a model for radar echo attenuation correction; the echo attenuation correction model is trained based on the following steps: acquiring the differential propagation phase shift rate samples and acquiring the attenuation rate samples; determining the network architecture based on the differential propagation phase shift rate samples; optimizing and training the network architecture based on the attenuation rate samples to obtain the echo attenuation correction model; the acquisition of the attenuation rate samples includes: acquiring... Take the initial first reflectivity factor set of the radar in the first band and the initial second reflectivity factor set of the radar in the second band; wherein the first band and the second band are different bands; determine the first reflectivity factor set and the second reflectivity factor set based on the initial first reflectivity factor set and the initial second reflectivity factor set; wherein the first reflectivity factor set is the factor set after preprocessing the initial first reflectivity factor set, and the second reflectivity factor set is the factor set after preprocessing the initial second reflectivity factor set; interpolate the second reflectivity factor set into the first reflectivity factor set to obtain the attenuation rate sample.

6. An electronic 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 computer program, it implements the radar echo attenuation correction method as described in any one of claims 1 to 4.

7. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the radar echo attenuation correction method as described in any one of claims 1 to 4.

8. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the radar echo attenuation correction method as described in any one of claims 1 to 4.

Images