Radar echo attenuation correction method and device, electronic equipment, medium and product

Abstract

The invention provides an echo attenuation correction method and device of a radar, electronic equipment, a medium and a product, and relates to the technical field of radars, and the method comprises the steps: obtaining the difference propagation phase shift rate of the radar; inputting the difference 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 a difference propagation phase shift rate sample and an attenuation rate sample, and the echo attenuation correction model is a model for performing echo attenuation correction of the radar. According to the technical scheme, the difference propagation phase shift rate of the radar is processed through the trained echo attenuation correction model, the echo attenuation correction result of the radar is obtained, the echo attenuation correction model has better stability and generalization ability, and the correction effect of the echo attenuation correction result of the radar is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of radar, and in particular, to a method, device, electronic device, medium and product for correcting echo attenuation of a radar. Background Art

[0002] Due to its advantages such as high spatio-temporal resolution, short wavelength, and high phase sensitivity, the X-band radar has stronger detection ability than conventional weather radars. Compared with ordinary Doppler weather radars, dual-polarization radars have significant advantages in short-term nowcasting, precipitation estimation, and phase state inversion. However, compared with C-band and S-band, the X-band radar has a higher frequency and more obvious attenuation in the atmosphere. How to effectively utilize these X-band radars, the most crucial is the attenuation correction of the radar.

[0003] Currently, in the prior art, traditional attenuation correction is mainly carried out based on empirical formulas. By substituting fixed empirical coefficients, parameters, etc. collected into the empirical formulas, the results of attenuation correction are obtained.

[0004] Therefore, when using traditional empirical formulas to correct the echo signal of the radar for attenuation, due to the use of fixed empirical coefficients, limited parameters, poor generalization ability, and great uncertainty, the correction results are inaccurate. Summary of the Invention

[0005] The present invention provides a method, device, electronic device, medium and product for correcting echo attenuation of a radar, aiming to solve the defect that in the prior art, when using traditional empirical formulas to correct the echo signal of the radar for attenuation, due to the use of fixed empirical coefficients, limited parameters, poor generalization ability, and great uncertainty, the correction results are inaccurate. It realizes processing the differential propagation phase shift rate of the radar through a trained echo attenuation correction model to obtain the echo attenuation correction result of the radar. The echo attenuation correction model has better stability and generalization ability to improve the correction effect of the echo attenuation correction result of the radar.

[0006] The present invention provides a method for correcting echo attenuation of a radar, including 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 correcting the echo attenuation of the radar.

[0009] A method for correcting echo attenuation of a radar 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 obtaining attenuation rate samples; determining a network architecture according to the differential propagation phase shift rate samples; optimizing and training the network architecture according to the attenuation rate samples to obtain an echo attenuation correction model.

[0010] A method for correcting echo attenuation of a radar provided by the present invention, obtaining attenuation rate samples includes: obtaining an initial first reflectivity factor set of a radar in a first band and an initial second reflectivity factor set of 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 according to the initial first reflectivity factor set and the initial second reflectivity factor set; wherein, the first reflectivity factor set is a factor set obtained by preprocessing the initial first reflectivity factor set, and the second reflectivity factor set is a factor set obtained by preprocessing the initial second reflectivity factor set; interpolating the second reflectivity factor set into the first reflectivity factor set to obtain attenuation rate samples.

[0011] A method for correcting echo attenuation of a radar provided by the present invention, determining a first reflectivity factor set and a second reflectivity factor set according to the initial first reflectivity factor set and the 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 the initial second reflectivity factor in the initial second reflectivity factor set; determining whether all the first correlation coefficients and all the second correlation coefficients are less than a preset coefficient; removing the initial first reflectivity factors corresponding to the first correlation coefficients less than the preset coefficient 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 coefficient from the initial second reflectivity factor set to obtain a candidate second reflectivity factor set; performing data normalization processing on all the initial first reflectivity factors in the candidate first reflectivity factor set to obtain a first reflectivity factor set, and performing data normalization processing on all the initial second reflectivity factors in the candidate second reflectivity factor set to obtain a second reflectivity factor set.

[0012] A method for correcting echo attenuation of a radar provided by the present invention determines a network architecture according to differential propagation phase shift rate samples, including: inputting the differential propagation phase shift rate samples into an encoder of a basic architecture to obtain an encoding result output by the encoder of the basic architecture; wherein, the basic architecture is a preset basic network architecture; inputting the encoding result into a decoder of the basic architecture to obtain a decoding result output by the decoder of the basic architecture; inputting the decoding result into a fully connected layer of the basic architecture to obtain an attenuation result output by the fully connected layer of the basic architecture; further optimizing and training the basic architecture according to the attenuation result to determine the network architecture.

[0013] A method for correcting echo attenuation of a radar provided by the present invention, inputting the differential propagation phase shift rate samples into an encoder of a basic architecture to obtain an encoding result output by the encoder of the basic architecture, including: inputting the differential propagation phase shift rate samples into a linear encoding layer in the encoder to obtain an encoding feature output by the linear encoding layer; inputting the encoding feature into a position encoding layer in the encoder to obtain an encoding result output by the position encoding layer.

[0014] The present invention also provides a device for correcting echo attenuation of a radar, including the following modules.

[0015] An acquisition module, configured to acquire the differential propagation phase shift rate of the radar.

[0016] A correction module, configured to input 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 trained based on differential propagation phase shift rate samples and attenuation rate samples, and the echo attenuation correction model is a model for correcting echo attenuation of the radar.

[0017] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, and when the processor executes the computer program, it implements any one of the above-mentioned methods for correcting echo attenuation of a radar.

[0018] The present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements any one of the above-mentioned methods for correcting echo attenuation of a radar.

[0019] The present invention also provides a computer program product, including a computer program, and when the computer program is executed by a processor, it implements any one of the above-mentioned methods for correcting echo attenuation of a radar.

[0020] A method, apparatus, electronic device, medium and product for correcting echo attenuation of a radar provided by the present invention obtain the differential propagation phase shift rate of the radar; input 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 trained based on a differential propagation phase shift rate sample and an attenuation rate sample, and the echo attenuation correction model is a model for correcting the echo attenuation of the radar. The technical solution of the present invention is used to solve the defect that in the prior art, when using a traditional empirical formula to correct the attenuation of a radar echo signal, due to the use of fixed empirical coefficients, limited parameters, poor generalization ability, and great uncertainty, the correction result is inaccurate. It realizes processing the differential propagation phase shift rate of the radar through a trained echo attenuation correction model to obtain the echo attenuation correction result of the radar. The echo attenuation correction model has better stability and generalization ability to improve the correction effect of the echo attenuation correction result of the radar. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0022] Figure 1 It is a flowchart showing the method for correcting echo attenuation of the radar provided by the present invention.

[0023] Figure 2 It is a structural diagram showing the apparatus for correcting echo attenuation of the radar provided by the present invention.

[0024] Figure 3 It is a structural diagram showing the electronic device provided by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0025] To make the objectives, technical solutions, and advantages of the present invention clearer, the following will clearly and completely describe the technical solutions in the present invention in conjunction with the drawings in the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art without creative efforts based on the embodiments in the present invention fall within the protection scope of the present invention.

[0026] The following will be combined with Figure 1The echo attenuation correction method of the radar provided by the present invention is described. The echo attenuation correction method of the radar provided by the present invention 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 an echo attenuation correction device of the radar set in the electronic device. The echo attenuation correction device of the radar can be implemented by software, hardware, or a combination of both. Figure 1 is a schematic flowchart of the echo attenuation correction method of the radar provided by the present invention. As Figure 1 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, and this embodiment does not limit this.

[0029] The differential propagation phase shift rate is a parameter used in meteorological radars 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 the horizontally and vertically polarized waves on the propagation path per unit distance. For example, it can be the change rates of three elevation angle layers of 1.5°, 2.4°, and 3.4° of the radar, and this embodiment does not limit this.

[0030] Specifically, obtain the differential propagation phase of the radar within a unit distance, and obtain the differential propagation phase shift rate of the radar according to the ratio of the differential propagation phase of the radar to the unit distance.

[0031] Step 102: 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.

[0032] In this step, 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 of the radar.

[0033] The attenuation rate samples can be, for example, horizontal attenuation rate samples, differential attenuation rate samples, etc. The echo attenuation correction results can be, for example, horizontal reflectivity factor attenuation correction results, differential reflectivity factor attenuation correction results, and this embodiment does not limit this. It can be trained based on differential propagation phase shift rate samples and horizontal attenuation rate samples so that the obtained echo attenuation correction model can analyze the input differential propagation phase shift rate to obtain the horizontal reflectivity factor attenuation correction result; it can also be trained based on differential propagation phase shift rate samples and differential attenuation rate samples so that the obtained echo attenuation correction model can analyze the input differential propagation phase shift rate to obtain 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, and then the echo attenuation correction model performs training processing on the input differential propagation phase shift rate, so as to obtain the echo attenuation correction result output by the echo attenuation correction model.

[0035] Exemplarily, when it is determined that the attenuation rate sample is a horizontal attenuation rate sample, the differential propagation phase shift rate sample and the horizontal attenuation rate sample can be used as the input of the echo attenuation correction model for model training, so that the obtained echo attenuation correction model processes the differential propagation phase shift rate to obtain the horizontal reflectivity factor attenuation correction result output by the echo attenuation correction model; when it is determined that the attenuation rate sample is a differential attenuation rate sample, the differential propagation phase shift rate sample and the differential attenuation rate sample can be used as the input of the echo attenuation correction model for model training, so that the obtained echo attenuation correction model processes the differential propagation phase shift rate to obtain the differential reflectivity factor attenuation correction result output by the echo attenuation correction model. Different attenuation rate samples such as horizontal attenuation rate samples and differential attenuation rate samples are used to train the model, so that the output results obtained by the final echo attenuation correction model processing the differential propagation phase shift rate are also different, and the obtained output results are related to different attenuation rate samples such as horizontal attenuation rate samples and differential attenuation rate samples.

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

[0037] In this step, the differential propagation phase shift rate sample is a sample for network construction. The attenuation rate sample is a sample for optimizing and training the constructed network architecture.

[0038] Specifically, obtain a differential propagation phase shift rate sample, determine a network architecture according to the differential propagation phase shift rate sample, obtain an attenuation rate sample, and optimize and train the network architecture according to the attenuation rate sample to obtain the echo attenuation correction model.

[0039] In a specific embodiment, obtaining the attenuation rate sample includes: obtaining 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; determining the first reflectivity factor set and the second reflectivity factor set according to 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; 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 can be, for example, the X band of the dual-polarization radar, and the second band can be, for example, the S band of the new generation weather radar. This embodiment does not limit this.

[0041] The initial first reflectivity factor set is the set of first reflectivity factors on different range bins of the X band at 150 kilometers of the dual-polarization radar. The initial first reflectivity factor set can 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 bin refers to the small unit divided by distance along the ray direction in the radar echo signal processing. The number of range bins can be, for example, 15. This embodiment does not limit this.

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

[0044] The preprocessing can be, for example, performing quality control, normalization, etc. on the initial first reflectivity factor set and the initial second reflectivity factor set respectively. The quality control can be, for example, filtering non-meteorological echoes, ground objects, and isolated clutter, etc. This embodiment does not limit this.

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

[0046] (1) In formula (1), represents the number of the range bin, for example, it can be 1, 2, 3…, represents the collected distance, for example, it can be 1.125 kilometers, represents the second horizontal reflectivity factor of a certain range bin in the S-band of a new generation weather radar, and represents the first horizontal reflectivity factor of a certain range bin in the X-band of a dual-polarization radar.

[0047] Differential attenuation rate sample is calculated as shown in formula (2).

[0048] (2) In formula (2), represents the number of the range bin, which can be 1, 2, 3..., for example, represents the collected range, which can be 1.125 kilometers, for example, represents the second differential reflectivity factor of a certain range bin in the S-band of a new generation weather radar, and represents the first differential reflectivity factor of a certain range bin in the X-band of a dual-polarization radar, which is not limited in this embodiment.

[0049] In a specific embodiment, obtaining the attenuation rate sample includes: obtaining 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; where the first band and the second band are different bands; determining the first reflectivity factor set and the second reflectivity factor set according to the initial first reflectivity factor set and the initial second reflectivity factor set; where 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; interpolating the second reflectivity factor set into the first reflectivity factor set to obtain the attenuation rate sample.

[0050] Specifically, in the first band, all initial first reflectivity factors of the radar at different range bins are obtained, so as to determine the first reflectivity factor set according to all the first reflectivity factors. In the second band, all initial second reflectivity factors of the radar at different range bins are obtained, so as to determine the second reflectivity factor set according to all the second reflectivity factors, and then quality control, normalization and other preprocessing are respectively performed on the initial first reflectivity factor set and the initial second reflectivity factor set to obtain the preprocessed first reflectivity factor set and the second reflectivity factor set; finally, the preprocessed second reflectivity factor set is interpolated into the preprocessed first reflectivity factor set by the method of bilinear interpolation to obtain the attenuation rate sample.

[0051] In a specific embodiment, determining the first set of reflectivity factors and the second set of reflectivity factors according to the initial first set of reflectivity factors and the initial second set of reflectivity factors includes: obtaining all the first correlation coefficients corresponding to all the initial first reflectivity factors in the initial first set of reflectivity factors, and determining all the second correlation coefficients corresponding to the initial second reflectivity factors in the initial second set of reflectivity factors; determining whether all the first correlation coefficients and all the second correlation coefficients are less than a preset coefficient; removing the initial first reflectivity factors corresponding to the first correlation coefficients less than the preset coefficient from the initial first set of reflectivity factors to obtain a candidate first set of reflectivity factors, and removing the initial second reflectivity factors corresponding to the second correlation coefficients less than the preset coefficient from the initial second set of reflectivity factors to obtain a candidate second set of reflectivity factors; performing data normalization processing on all the initial first reflectivity factors in the candidate first set of reflectivity factors to obtain the first set of reflectivity factors, and performing data normalization processing on all the initial second reflectivity factors in the candidate second set of reflectivity factors to obtain the second set of reflectivity factors.

[0052] In this step, both the first correlation coefficient and the second correlation coefficient can represent the correlation between the horizontally polarized echo and the vertically polarized echo within the radar resolution volume, and the value range is [0, 1]. The closer the value is to 1, the better the data quality of all the initial first reflectivity factors in the initial first set of reflectivity factors and all the initial second reflectivity factors in the initial second set of reflectivity factors. The first correlation coefficient and the second correlation coefficient can be used to judge the data quality of all the initial first reflectivity factors in the initial first set of reflectivity factors and all the initial second reflectivity factors in the initial second set of reflectivity factors. Low first and second correlation coefficients may mean that there are non-meteorological echoes or noises in all the initial first reflectivity factors in the initial first set of reflectivity factors and all the initial second reflectivity factors in the initial second set of reflectivity factors, indicating the existence of non-uniform phase states or noises in the corresponding range bin areas.

[0053] For example, the first correlation coefficient can be 0.95, and the second correlation coefficient can be 0.95. This embodiment does not limit this.

[0054] The preset coefficient is a coefficient preset for judging the first correlation coefficient and the second correlation coefficient. For example, it can be 0.9. This embodiment does not limit this.

[0055] Specifically, obtain all the first correlation coefficients corresponding to all the initial first reflectivity factors in the initial first reflectivity factor set, and determine all the second correlation coefficients corresponding to the initial second reflectivity factor 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; remove the initial first reflectivity factors corresponding to the first correlation coefficients less than the preset coefficient from the initial first reflectivity factor set to obtain a candidate first reflectivity factor set, and remove the initial second reflectivity factors corresponding to the second correlation coefficients less than the preset coefficient from the initial second reflectivity factor set to obtain a candidate second reflectivity factor set; perform data unification processing on all the initial first reflectivity factors in the candidate first reflectivity factor set to obtain a first reflectivity factor set, and perform data unification processing on all the initial second reflectivity factors in the candidate second reflectivity factor set to obtain a second reflectivity factor set.

[0056] Exemplarily, the processing method for performing data unification processing on all the initial first reflectivity factors in the candidate first reflectivity factor set may be, for example, normalization processing according to the maximum and minimum values of the -20 to 70 decibel reflectivity factor (dBZ); the processing method for performing data unification processing on all the initial second reflectivity factors in the candidate second reflectivity factor set may be, for example, normalization processing according to the maximum and minimum values of -2 to 6 dB.

[0057] In a specific embodiment, determining the network architecture according to the differential propagation phase shift rate samples includes: inputting the differential propagation phase shift rate samples into the encoder of the basic architecture to obtain the encoding result output by the encoder of the basic architecture; wherein, the basic architecture is a preset basic network architecture; inputting the encoding result into the decoder of the basic architecture to obtain the decoding result output by the decoder of the basic architecture; inputting the decoding result into the fully connected layer of the basic architecture to obtain the attenuation result output by the fully connected layer of the basic architecture; further optimizing and training the basic architecture according to the attenuation result to determine the network architecture.

[0058] In this step, before determining the network architecture, data unification settings can also be performed on the differential propagation phase shift rate samples to improve the accuracy of subsequent model training. The basic architecture is a preset untrained basic network architecture, and the basic architecture includes an encoder, a decoder, and a fully connected layer, which are not limited in this embodiment.

[0059] The basic architecture includes an encoder, a decoder, and finally outputs the echo attenuation correction result through a fully connected layer. The basic architecture may be, for example, an X-band radar echo attenuation correction framework designed based on Transformer (a model architecture), which is not limited in this embodiment.

[0060] A masking mechanism is introduced into the decoder to improve the robustness of the model. Then, through the parallel computing of the multi-head attention mechanism (for example, it can be 8 heads), the model can simultaneously focus on different parts of the sequence of the input differential propagation phase shift rate samples, learn information from multiple subspaces, and add residual connections and normalization to improve the convergence speed and promote better feature learning. Then it is passed to the feed-forward neural network for further non-linear mapping and feature extraction, and two decoding modules (2×) are set for cyclic decoding. Finally, the decoding result is input into the fully connected layer to obtain the output attenuation result.

[0061] Specifically, input the differential propagation phase shift rate samples into the encoder of the basic architecture to obtain the encoded result output by the encoder of the basic architecture; input the encoded result into the decoder of the basic architecture to obtain the decoded result output by the decoder of the basic architecture; input the decoded result into the fully connected layer of the basic architecture to obtain the attenuation result output by the fully connected layer of the basic architecture; further optimize and train the basic architecture according to the attenuation result to determine the network architecture.

[0062] In a specific embodiment, inputting the differential propagation phase shift rate samples into the encoder of the basic architecture to obtain the encoded result output by the encoder of the basic architecture includes: inputting the differential propagation phase shift rate samples into the linear encoding layer in the encoder to obtain the encoded features output by the linear encoding layer; inputting the encoded features into the position encoding layer in the encoder to obtain the encoded result output by the position encoding layer.

[0063] In this step, the linear encoding layer can be, for example, a layer that performs linear encoding through 32-dimensional vectors and 64-dimensional vectors respectively. This embodiment does not limit this.

[0064] Specifically, input the differential propagation phase shift rate samples into the linear encoding layer in the encoder, encode through two layers of linear encoding layers to obtain the encoded features output by the linear encoding layer; input the encoded features into the position encoding layer in the encoder to supplement the position information through the position encoding layer to obtain the encoded result output by the position encoding layer.

[0065] Exemplarily, for instance, it can be to obtain the effective value of the differential propagation phase shift rate samples of 15 range bins as the input, encode it into a 64-dimensional vector through two linear layers in the encoder, and supplement the position information of the differential propagation phase shift rate samples through the position encoding layer to make the front and back related; then introduce a masking mechanism in the decoder to improve the robustness of the model, and then through the parallel calculation of the multi-head attention mechanism, enable the model to simultaneously focus on different parts in the sequence of the differential propagation phase shift rate samples, learn information from multiple subspaces, and add residual connections and normalization to improve the convergence speed and promote better feature learning, and then pass it to the feed-forward neural network for further non-linear mapping and feature extraction, and set two decoding modules (2×) for cyclic decoding steps, and finally input it into the fully connected layer to obtain the final attenuation result, and then obtain the echo attenuation correction result according to the attenuation result. The activation function all adopts the rectified linear unit (ReLU), and Dropout (random inactivation) is used to prevent overfitting, which is not limited in this embodiment.

[0066] In a specific embodiment, after obtaining the horizontal reflectivity factor attenuation correction result or the differential reflectivity factor attenuation correction result, it is also possible to further evaluate the horizontal reflectivity factor attenuation correction result or the differential reflectivity factor attenuation correction result to obtain an evaluation result.

[0067] Calculate evaluation indexes such as ratio bias, root mean square error and mean absolute error through the horizontal reflectivity factor attenuation correction result or the differential reflectivity factor attenuation correction result, and evaluate the horizontal reflectivity factor attenuation correction result or the differential reflectivity factor attenuation correction result through the ratio bias, root mean square error and mean absolute error.

[0068] Among them, the ratio bias is calculated as shown in formula (3), the root mean square error is calculated as shown in formula (4), and the mean absolute error is calculated as shown in formula (5).

[0069] (3) (4) (5) In formulas (3), (4) and (5), represents the second reflectivity factor corresponding to the th range bin in the second reflectivity factor set, represents the horizontal reflectivity factor attenuation correction result or the differential reflectivity factor attenuation correction result obtained after correction, Represents the total quantity of the distance library. BIAS can measure the difference between segment X and segment S. The closer its value is to 1, the smaller the difference in radar intensities of different sizes; RMSE can measure the dispersion degree of the intensity difference. The smaller its value, the smaller the difference in radar intensities of different sizes; MAE represents the average level of the absolute difference between echoes and is less affected by outliers. The smaller the value of MAE, the higher the absolute accuracy of the attenuation correction result of the horizontal reflectivity factor or the attenuation correction result of the differential reflectivity factor.

[0070] Exemplarily, as shown in Table 1, it shows the horizontal reflectivity factor / differential reflectivity factor before correction and the corrected horizontal reflectivity factor / differential reflectivity factor in the echo attenuation correction result after correction by the echo attenuation correction model, as well as the horizontal reflectivity factor / differential reflectivity factor corrected by the existing empirical formula for comparison. It is determined that the result of correction by the echo attenuation correction model has improved compared to the correction effect of the empirical formula, improving the correction accuracy.

[0071] Table 1

[0072] As shown in Table 1, by comparing the BIAS before and after the correction of the horizontal reflectivity factor, the model improves the BIAS from 0.875 to 0.972, which is better than 0.901 after correction by the empirical formula. For the differential reflectivity factor, the BIAS before correction is 0.862, 1.141 after the model, while it is 1.273 after correction by the empirical formula, showing an overcorrection problem. Before the correction of the horizontal reflectivity factor, RMSE and MAE are 8.693 and 6.292 respectively, and after correction by the model, they are 5.811 and 4.222 respectively, with improvements of 33.15% and 32.89% respectively. After correction by the empirical formula, they are 6.820 and 5.113 respectively, with improvements of 21.54% and 18.73% respectively. Before the correction of the differential reflectivity factor, RMSE and MAE are 1.679 and 1.271 respectively, and after correction by the model, they are 0.972 and 0.697 respectively, with improvements of 42.10% and 45.16% respectively. After correction by the empirical formula, they are 1.382 and 1.008 respectively, with improvements of 17.69% and 20.69% respectively. All evaluation indicators have obvious improvements compared to the empirical formula correction method.

[0073] A method for correcting echo attenuation of a radar provided by the present invention 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 trained based on differential propagation phase shift rate samples and attenuation rate samples, and the echo attenuation correction model is a model for correcting the echo attenuation of the radar. Based on the above embodiments, the technical solution of the present invention is used to solve the defect that in the prior art, when using a traditional empirical formula to correct the attenuation of the radar echo signal, due to the use of fixed empirical coefficients, limited parameters, poor generalization ability, and great uncertainty, the correction result is inaccurate. It realizes processing the differential propagation phase shift rate of the radar through the trained echo attenuation correction model to obtain the echo attenuation correction result of the radar. The echo attenuation correction model has better stability and generalization ability to improve the correction effect of the echo attenuation correction result of the radar.

[0074] 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 correspondingly referred to the radar echo attenuation correction method described above.

[0075] Figure 2 is a schematic structural diagram of the radar echo attenuation correction device provided by the present invention. Refer to Figure 2 As shown, the radar echo attenuation correction device 200 includes: an acquisition module 301 and a correction module 302.

[0076] The acquisition module 301 is used to acquire the differential propagation phase shift rate of the radar.

[0077] The correction module 302 is used to input the differential propagation phase shift rate into the 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 trained based on differential propagation phase shift rate samples and attenuation rate samples, and the echo attenuation correction model is a model for correcting the echo attenuation of the radar.

[0078] In an exemplary embodiment, the device further includes: a model training module. The model training module is used to: acquire differential propagation phase shift rate samples and acquire attenuation rate samples; determine a network architecture according to the differential propagation phase shift rate samples; optimize and train the network architecture according to the attenuation rate samples to obtain an echo attenuation correction model.

[0079] In an exemplary embodiment, the model training module obtains attenuation rate samples, specifically: obtaining an initial first reflectivity factor set of a radar in a first band and an initial second reflectivity factor set of 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 according to the initial first reflectivity factor set and the initial second reflectivity factor set; wherein, the first reflectivity factor set is a factor set obtained by preprocessing the initial first reflectivity factor set, and the second reflectivity factor set is a factor set obtained by 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.

[0080] In an exemplary embodiment, the model training module determines a first reflectivity factor set and a second reflectivity factor set according to the initial first reflectivity factor set and the initial second reflectivity factor set, specifically: 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 the first correlation coefficients and all the second correlation coefficients are less than a preset coefficient; removing the initial first reflectivity factors corresponding to the first correlation coefficients less than the preset coefficient 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 coefficient from the initial second reflectivity factor set to obtain a candidate second reflectivity factor set; performing data unification processing on all the 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 the initial second reflectivity factors in the candidate second reflectivity factor set to obtain a second reflectivity factor set.

[0081] In an exemplary embodiment, the model training module determines a network architecture according to the differential propagation phase shift rate samples, specifically: inputting the differential propagation phase shift rate samples into the encoder of the basic architecture to obtain an encoding result output by the encoder of the basic architecture; wherein, the basic architecture is a preset basic network architecture; inputting the encoding result into the decoder of the basic architecture to obtain a decoding result output by the decoder of the basic architecture; inputting the decoding result into the fully connected layer of the basic architecture to obtain an attenuation result output by the fully connected layer of the basic architecture; further optimizing and training the basic architecture according to the attenuation result to determine the network architecture.

[0082] In an exemplary embodiment, the model training module inputs the differential propagation phase shift rate samples into the encoder of the infrastructure to obtain the encoded result output by the encoder of the infrastructure. Specifically, it is used to: input the differential propagation phase shift rate samples into the linear encoding layer in the encoder to obtain the encoded features output by the linear encoding layer; input the encoded features into the position encoding layer in the encoder to obtain the encoded result output by the position encoding layer.

[0083] The device in this embodiment can be used to execute the method in any one of the method embodiments on the side of the radar echo attenuation correction method. Its specific implementation process and technical effects are similar to those in the method embodiments on the side of the radar echo attenuation correction method. For specific details, reference can be made to the detailed introduction in the method embodiments on the side of the radar echo attenuation correction method, and details will not be elaborated here.

[0084] Figure 3 is a schematic structural diagram of the electronic device provided by the present invention. As Figure 3 shown, the electronic device may include: a processor 310, a communication interface 320, a memory 330, and a communication bus 340. Among them, the processor 310, the communication interface 320, and the memory 330 communicate with each other through the communication bus 340. The processor 310 can call the logical instructions in the memory 330 to execute the radar echo attenuation correction method, which 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 trained based on the differential propagation phase shift rate samples and the attenuation rate samples, and the echo attenuation correction model is a model for performing radar echo attenuation correction.

[0085] In addition, when the logical instructions in the above-mentioned memory 330 are implemented in the form of software functional units and sold or used as an independent product, they 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 this 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 for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods in various embodiments of the present invention. The aforementioned storage medium includes: USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical disks, and other various media that can store program codes.

[0086] On the other hand, the present invention also provides a computer program product. The computer program product 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 echo attenuation correction method of the radar provided by the above-mentioned various 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 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 the radar.

[0087] In another aspect, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it is implemented to execute the echo attenuation correction method of the radar provided by the above-mentioned various 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 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 the radar.

[0088] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative efforts.

[0089] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the above technical solution, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disc, etc., and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0090] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for correcting echo attenuation of a radar, characterized in that, Including: 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 trained based on a differential propagation phase shift rate sample and an attenuation rate sample, and the echo attenuation correction model is a model for correcting the echo attenuation of the radar.

2. The echo attenuation correction method of the radar according to claim 1, wherein The echo attenuation correction model is trained based on the following steps: Obtaining the differential propagation phase shift rate sample and obtaining the attenuation rate sample; Determining a network architecture according to the differential propagation phase shift rate sample; Optimally training the network architecture according to the attenuation rate sample to obtain the echo attenuation correction model.

3. The echo attenuation correction method of the radar according to claim 2, characterized in that, The obtaining the attenuation rate sample includes: Obtaining an initial first reflectivity factor set of a radar in a first band and an initial second reflectivity factor set of 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 according to the initial first reflectivity factor set and the initial second reflectivity factor set; wherein, the first reflectivity factor set is a factor set obtained by preprocessing the initial first reflectivity factor set, and the second reflectivity factor set is a factor set obtained by 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.

4. The method for correcting echo attenuation of the radar according to claim 3, characterized in that, The determining the first reflectivity factor set and the second reflectivity factor set according to the initial first reflectivity factor set and the 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 the initial second reflectivity factors in the initial second reflectivity factor set; Determining whether all the first correlation coefficients and all the second correlation coefficients are less than a preset coefficient; Excluding the initial first reflectivity factors corresponding to the first correlation coefficients less than the preset coefficient from the initial first reflectivity factor set to obtain a candidate first reflectivity factor set, and excluding the initial second reflectivity factors corresponding to the second correlation coefficients less than the preset coefficient from the initial second reflectivity factor set to obtain a candidate second reflectivity factor set; Performing data unification processing on all the initial first reflectivity factors in the candidate first reflectivity factor set to obtain the first reflectivity factor set, and performing data unification processing on all the initial second reflectivity factors in the candidate second reflectivity factor set to obtain the second reflectivity factor set.

5. The echo attenuation correction method of the radar according to claim 2, characterized in that, The determining a network architecture according to the differential propagation phase shift rate sample includes: Inputting the differential propagation phase shift rate sample into an encoder of a basic architecture to obtain a coding result output by the encoder of the basic architecture; wherein, the basic architecture is a preset 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; Further optimize and train the infrastructure according to the attenuation result to determine the network architecture.

6. The echo attenuation correction method for a radar according to claim 5, 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: Input the differential propagation phase shift rate sample into the linear encoding layer in the encoder to obtain the encoding features output by the linear encoding layer; Input the encoding features into the position encoding layer in the encoder to obtain the encoding result output by the position encoding layer.

7. An echo attenuation correction device for a radar, characterized in that, It includes: An acquisition module for acquiring the differential propagation phase shift rate of the radar; A correction module for 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 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 the radar.

8. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the echo attenuation correction method of the radar according to any one of claims 1 to 6.

9. 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 echo attenuation correction method of the radar according to any one of claims 1 to 6.

10. A computer program product comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the echo attenuation correction method of the radar according to any one of claims 1 to 6.

Images