Interference suppression method based on distance-azimuth two-dimensional interception processing

Abstract

The invention discloses an interference suppression method based on distance-azimuth two-dimensional interception processing, and the method comprises the steps: carrying out the distance-azimuth two-dimensional interception of a whole SLC image containing interference, and obtaining an intercepted sub-image; wherein a protection unit is set on the basis of a preset distance dimension interception point number when distance dimension interception is carried out; performing Fourier transform on each sub-image along a distance dimension to obtain a two-dimensional distance spectrum of the sub-image, performing pulse-by-pulse outlier detection, and constructing a two-dimensional interference mask of the sub-image based on the outliers; carrying out frequency domain wave trapping by utilizing the two-dimensional interference mask of the sub-image, carrying out point multiplication operation on the two-dimensional interference mask and the two-dimensional distance frequency spectrum of the sub-image so that a frequency point with an interference component is processed to be zero, and then obtaining a sub-image distance frequency spectrum after interference suppression processing; performing inverse Fourier transform on the sub-image distance spectrum along the distance dimension to obtain a sub-image after interference suppression; and constructing a whole SLC image after interference suppression based on the sub-images after interference suppression.

Description

Technical Field

[0001] The present invention relates to the field of radar signal processing, and particularly to an interference suppression method based on range-azimuth two-dimensional intercept processing. Background Art

[0002] Radar remote sensing satellites can obtain two-dimensional high-resolution images all day and all weather, and have unique advantages in military and civilian applications. They are an important part of China's high-resolution earth observation system. However, radar remote sensing satellites are extremely vulnerable to interference from other electromagnetic radiation sources in the same frequency band. The interference signal will form masking artifacts in the imaging result, seriously affecting the image quality, possibly leading to missed detection and false alarms of targets, and greatly affecting the reliability of subsequent remote sensing application qualitative interpretation.

[0003] At present, many suppression ideas have been proposed for narrowband interference and broadband interference, but these methods are not necessarily applicable to complex interference scenarios where there are multiple signals with different characteristics in the same scene. When analyzing measured data, it is found that in complex interference scenarios, narrowband interference and broadband interference may exist simultaneously. If the narrowband interference and broadband interference are processed separately, the processing efficiency is too low and the computational cost is large; currently, there is no method that can efficiently and quickly process complex interference. Summary of the Invention

[0004] The object of the present invention is to provide an interference suppression method based on range-azimuth two-dimensional intercept processing to overcome the problems existing in the prior art.

[0005] To achieve the above task, the present invention adopts the following technical solutions:

[0006] An interference suppression method based on range-azimuth two-dimensional intercept processing, comprising:

[0007] Performing range-azimuth two-dimensional intercept on the entire SLC image containing interference to obtain the intercepted sub-image; wherein when performing the intercept in the range dimension, a protection unit is set on the basis of the preset number of range dimension intercept points to change the size of the sub-image when intercepted in the range dimension;

[0008] Performing Fourier transform on each intercepted sub-image along the range dimension to obtain the two-dimensional range spectrum of the sub-image; then detecting outliers along the range dimension for the two-dimensional range spectrum, and the detected outliers are the frequency points where interference components exist, and constructing a two-dimensional interference mask of the sub-image based on the outliers;

[0009] Perform frequency-domain notch filtering using the two-dimensional interference mask of the sub-image. Multiply the two-dimensional interference mask with the two-dimensional range spectrum of the sub-image to set the frequency points with interference components to zero, and then obtain the range spectrum of the sub-image after interference suppression processing. Perform inverse Fourier transform on this sub-image range spectrum along the range dimension to obtain the sub-image after interference suppression.

[0010] Construct the entire SLC image after interference suppression based on the sub-image after interference suppression.

[0011] Further, the two-dimensional range-azimuth truncation of the entire SLC image with interference to obtain the truncated sub-image includes:

[0012] For the entire SLC image x with interference of size Na×Nr img (t a , t r ) perform two-dimensional range-azimuth truncation, where N r and N a respectively represent the number of sampling points of the SLC image in the range dimension and the azimuth dimension, t a represents the azimuth sampling time, t r represents the range sampling time. Set the number of guard cells in the range dimension to N pro , the number of truncated points in the azimuth dimension of the sub-image is M, and the number of truncated points in the range dimension of the sub-image is N; the two-dimensional range-azimuth truncation is an operation of truncating the sub-image of the SLC image along the range dimension and the azimuth dimension. Among them, setting the range guard cell means taking N pro more cells on both sides during range dimension truncation, that is, the number of truncated points in the range dimension of the first sub-image and the last sub-image is (N + N pro ), and the number of truncated points in the range dimension of the remaining sub-images is (N + 2 * N pro ).

[0013] Further, use the azimuth-dimensional rectangular window function and the range-dimensional rectangular window function to perform two-dimensional range-azimuth truncation; the i-th sub-image x sub_i (t a , t r ), i = 1, …, N a N r / MN's mathematical expression is:

[0014]

[0015] Among them, the expression of the azimuth-dimensional rectangular window function is:

[0016]

[0017] The expression of the range - dimension rectangular window function is:

[0018]

[0019] After two - dimensional range - azimuth truncation, (N a N r / MN - 2*N a / M) sub - images with a size of M×(N + 2*N pro ) and 2*N a / M sub - images with a size of M×(N + N pro ) are obtained xsub_i ( ta,tr ).

[0020] Furthermore, the step of detecting outliers for each pulse along the range dimension of the two - dimensional range spectrum includes:

[0021] The judgment formula for outliers is:

[0022]

[0023] In the formula, represents the energy value of the two - dimensional range spectrum of the sub - image at the range frequency at the azimuth sampling time ; represents the spectrum corresponding to all range frequencies of the two - dimensional range spectrum X (t sub_i (t a , f r ) at the t a1 -th azimuth sampling time, median represents the median operation, and c is a constant;

[0024] When the above formula holds, the at this time is an outlier, that is, a frequency point where interference may exist.

[0025] Furthermore, the step of constructing a two - dimensional interference mask for the sub - image based on the outliers includes:

[0026] After detecting outliers for each azimuth sampling time of the two - dimensional range spectrum of the sub - image along the range dimension respectively, a two - dimensional interference mask mask sub_i of the sub - image is constructed; where the interference mask value at the azimuth sampling time and range frequency is calculated by the formula:

[0027]

[0028] Among them, is the mask indication quantity, represents at the azimuth sampling time The two-dimensional range spectrum corresponding to all range frequencies above.

[0029] Furthermore, the sub-image range spectrum after the interference suppression process is specifically expressed as:

[0030] Sub-image range spectrum X sub_i_after (t a , f r ) has the following calculation formula:

[0031] X sub_i_after (t a , f r ) = X sub_i (t a , f r ). * (1 - mask sub_i ).

[0032] Furthermore, constructing the entire SLC image after interference suppression based on the sub-image after interference suppression includes:

[0033] For the sub-image x sub_i_after (t a , t r ) after interference suppression, cut off the protection unit, that is, perform a two-dimensional cut of range - azimuth, for ([[]] N a N r / MN - 2 * N a / M ) sub-images after interference suppression with a size of M×(N + 2*N pro ) and ([[]] 2 * N a / M ) sub-images after interference suppression with a size of M×(N + N pro ) respectively cut off the corresponding protection units along the range dimension, obtaining (N a N r / MN) sub-images after interference suppression with a size of M×N, x' sub_i_after (t a , t r );Then splice the sub-images in the order of interception to obtain the entire SLC image after interference suppression.

[0034] A terminal device, comprising a processor, a memory, and a computer program stored in the memory; when the processor executes the computer program, the interference suppression method based on distance-azimuth two-dimensional intercept processing is implemented.

[0035] A computer-readable storage medium, in which a computer program is stored; when the computer program is executed by a processor, the interference suppression method based on distance-azimuth two-dimensional intercept processing is implemented.

[0036] Compared with the prior art, the present invention has the following technical features:

[0037] Compared with the conventional image domain interference suppression methods, the present invention performs sub-aperture decomposition on the entire SLC image along the range dimension to obtain multiple sub-images. Then, by analyzing the frequency feature differences between the useful signals and interference signals at different range times in each sub-image, various interferences with different energy intensities and different sources existing in the same scene can be quickly detected and suppressed, and the ground object information covered by interference artifacts in the scene can be restored. The interference suppression processing of the present invention has high efficiency and is easy to be implemented in engineering. Description of the Drawings

[0038] Figure 1 is a schematic flow chart for implementing the method of the present invention;

[0039] Figure 2 is the original SLC image with interference in the embodiment of the present invention, where the yellow area is interference artifacts;

[0040] Figure 3 is in the embodiment of the present invention Figure 2 After performing distance-azimuth two-dimensional intercept processing, partial sub-images with interference and the corresponding two-dimensional range spectra are obtained, where (a), (c), and (e) are sub-images without interference, and (b), (d), and (f) are two-dimensional range spectra corresponding to the sub-images with interference;

[0041] Figure 4 is the sub-image after interference suppression processing in the embodiment of the present invention, where Figure 4 of (a) is Figure 3 the result after interference suppression corresponding to (a) of Figure 4 of (b) is Figure 3 the result after interference suppression corresponding to (c) of Figure 4 of (c) is Figure 3 the result after interference suppression corresponding to (e) of

[0042] Figure 5 is the entire scene spaceborne SLC image after interference suppression processing in the embodiment of the present invention. Detailed Embodiment

[0043] The present invention provides an interference suppression method based on range-azimuth two-dimensional truncation processing. By performing range-azimuth two-dimensional truncation processing on the entire SLC image containing interference, the image is divided into multiple sub-image slices. Each sub-image only reflects the image features within a certain range time period. Similarly, the range spectrum of the sub-image also reflects the local frequency characteristics within this range time period. In the range spectrum of this sub-image, the energy of the interference signal is more concentrated. By combining the outlier detection of the energy outliers in the range spectrum of the sub-image and performing a zeroing operation, and then performing an inverse Fourier transform (IFFT) along the range dimension, the sub-image after interference suppression can be obtained. Then, the sub-images after interference suppression processing are spliced in the original order, and the entire single-channel complex image after interference suppression can be obtained. By analyzing the range frequency characteristics in different sub-images, the present invention enhances the feature differences between complex interference and useful signals, realizes effective detection and suppression of interference, is applicable to processing complex scenarios with multiple coexisting interference signals, and is efficient in processing and easy to implement in engineering. The specific implementation method of the present invention is as follows:

[0044] Step 1: Perform range-azimuth two-dimensional truncation on the entire SLC image x of size Na×Nr img (t a ,t r ), where N r and N a respectively represent the number of sampling points of the SLC image in the range dimension and the azimuth dimension, t a represents the azimuth sampling time, t r represents the range sampling time. Set the number of guard cells in the range dimension to N pro , the number of azimuth dimension truncation points of the sub-image is M, and the number of range dimension truncation points of the sub-image is N.

[0045] Range-azimuth two-dimensional truncation is an operation of truncating sub-images of the SLC image along the range dimension and the azimuth dimension. Among them, setting the range guard cell means taking N pro more cells on both sides during range dimension truncation, that is, the number of range dimension truncation points of the first sub-image and the last sub-image is (N + N pro ), and the number of range dimension truncation points of the remaining sub-images is (N + 2*N pro ); The range-azimuth two-dimensional truncation can be represented by the azimuth dimension rectangular window function and the range dimension rectangular window function ; The mathematical expression of the i-th sub-image x sub_i (t a ,t r ), i = 1, …, N a N r / MN is:

[0046]

[0047] Among them, the expression of the azimuth - dimension rectangular window function is:

[0048]

[0049] Because a guard cell is set in the range dimension, the expression of the range - dimension rectangular window function is divided into three cases:

[0050]

[0051] After range - azimuth two - dimensional truncation, we get (N a N r / MN - 2*N a / M) sub - images of size M×(N + 2*N pro ) and 2*N a / M sub - images of size M×(N + N pro ) xsub_i ( ta,tr ).

[0052] Step 2: Perform Fourier transform on each intercepted sub - image x sub_i (t a , t r ) along the range dimension to obtain the two - dimensional range spectrum X sub_i (t a , f r ) of the sub - image; then perform per - pulse detection of outliers on the two - dimensional range spectrum X sub_i (t a , f r ) along the range dimension. The detected outliers are the frequency points with interference components, and a two - dimensional interference mask mask sub_i of the sub - image is constructed based on the outliers.

[0053] The calculation formula for performing Fourier transform in the range dimension is:

[0054]

[0055] Among them, f r represents the range frequency, e represents the natural constant, and j is the imaginary unit.

[0056] Then perform per - pulse detection of outliers on the two - dimensional range spectrum along the range dimension. The detected outliers are the frequency points with interference components; the judgment formula for outliers is:

[0057]

[0058] In the formula, represents the range frequency at the azimuth sampling time Two-dimensional range spectrum of the sub-image at The energy value of represents the two-dimensional range spectrum X of the sub-image sub_i (t a , f r ) at the th azimuth sampling time corresponding to all range frequencies. Median represents the median operation, and c is a constant; in this embodiment, the empirical value c = 1.4826 is adopted.

[0059] When the above equation holds, the at this time is an outlier, that is, a frequency point where interference may exist.

[0060] After performing outlier detection on all azimuth sampling times of the two-dimensional range spectrum of the sub-image along the range dimension, a two-dimensional interference mask mask of the sub-image is constructed sub_i ; where the azimuth sampling time range frequency The interference mask value at The calculation formula is:

[0061]

[0062] Among them, is the mask indication quantity, represents the two-dimensional range spectrum corresponding to all range frequencies at the azimuth sampling time .

[0063] Step 3, use the two-dimensional interference mask mask of the sub-image sub_i to perform frequency domain notch filtering. Multiply the two-dimensional interference mask masksub_i with the two-dimensional range spectrum X of the sub-image sub_i (t a , f r ) to set the frequency points with interference components to zero, and then obtain the sub-image range spectrum X after interference suppression processing sub_i_after (t a , f r ); Perform inverse Fourier transform on this sub-image range spectrum X sub_i_after (t a , f r ) along the range dimension to obtain the sub-image x after interference suppression sub_i_after (t a , f r ).

[0064] Among them, the calculation formula of the sub-image range spectrum X sub_i_after (t a , f r ) is:

[0065] Xsub_i_after (t a , f r ) = X sub_i (t a , f r ).*(1 - mask sub_i );

[0066] Then, perform an inverse Fourier transform on the sub - image range spectrum X i_after (t a , f r ) along the range dimension to obtain ([[]] N a N r / MN - 2*N a / M ) sub - images after interference suppression with a size of M×(N + 2*N pro ) and ([[]] 2*N a / M ) sub - images after interference suppression with a size of M×(N + N pro ).

[0067] The inverse transform formula for the range dimension is:

[0068]

[0069] Step 4, truncate the protection unit from the sub - image x sub_i_after (t a , t r ). That is, referring to Step 1, for ([[]] N a N r / MN - 2 * N a / M ) sub - images after interference suppression with a size of M×(N + 2*N pro ) and ([[]] 2 * N a / M ) sub - images after interference suppression with a size of M×(N + N pro ), truncate the corresponding protection units along the range dimension respectively to obtain (N a N r / MN) sub - images after interference suppression with a size of M×N, x' sub_i_after (t a,t r ); Then splice the sub-images again according to the interception order in Step 1 to obtain the entire SLC image after interference suppression.

[0070] Example:

[0071] In one embodiment of the present invention, refer to Figures 2 to 4 , set the distance protection unit N pro = 10, the number of interception points in the azimuth dimension is M = 500, the number of interception points in the distance dimension is N = 500, for the SLC image x(t a ,t r ) with interference and a size of 4000×24000, perform two-dimensional distance-azimuth interception processing to obtain 16 sub-images x with a size of 500×510 and 352 sub-images x with a size of 500×520 sub_i (t a ,t r ) ， where i = 1,..., 368; finally, after interference suppression, 358 sub-images x' with a size of 500×500 are obtained sub_i_after (t a ,t r ); Then splice the sub-images again according to the slice splitting order in Step 1 to obtain the SLC image after interference suppression, as Figure 5 shown.

[0072] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it; although the present application 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 for 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 application, and should all be included in the protection scope of the present application.

Claims

1. A method for interference suppression based on distance-azimuth two-dimensional intercept processing, characterized in that, Including: Performing two-dimensional range-azimuth truncation on the entire SLC image with interference to obtain the truncated sub-image; Among them, when performing range dimension truncation, a protection unit is set based on the preset number of range dimension truncation points to change the size of the sub-image when truncated in the range dimension; Performing Fourier transform on each truncated sub-image along the range dimension to obtain the two-dimensional range spectrum of the sub-image; then detecting outliers pulse by pulse along the range dimension of the two-dimensional range spectrum, and the detected outliers are the frequency points with interference components, and constructing a two-dimensional interference mask of the sub-image based on the outliers; Performing frequency domain notch filtering using the two-dimensional interference mask of the sub-image, performing a point multiplication operation on the two-dimensional interference mask and the two-dimensional range spectrum of the sub-image to set the frequency points with interference components to zero, and then obtaining the sub-image range spectrum after interference suppression processing; performing inverse Fourier transform on this sub-image range spectrum along the range dimension to obtain the sub-image after interference suppression; Constructing the entire SLC image after interference suppression based on the sub-image after interference suppression.

2. The interference suppression method based on distance-azimuth two-dimensional intercept processing according to claim 1, wherein Performing two-dimensional range-azimuth truncation on the entire SLC image with interference to obtain the truncated sub-image, including: For the entire SLC image x with interference of size N a ×N r perform range-azimuth two-dimensional truncation, where N img (t a , t r ), N r and N a respectively represent the number of sampling points of the SLC image in the range dimension and the azimuth dimension, t a represents the azimuth sampling time, and t r represents the range sampling time. Set the number of guard cells in the range dimension to N pro . The number of truncated points in the azimuth dimension of the sub-image is M, and the number of truncated points in the range dimension of the sub-image is N; the range-azimuth two-dimensional truncation is an operation of truncating the sub-image of the SLC image along the range dimension and the azimuth dimension. Among them, setting the range guard cell means taking N pro more cells on each side during range dimension truncation, that is, the number of truncated points in the range dimension of the first sub-image and the last sub-image is (N + N pro ), and the number of truncated points in the range dimension of the remaining sub-images is (N + 2 * N pro ).

3. The interference suppression method based on distance-azimuth two-dimensional intercept processing according to claim 2, wherein Adopt an azimuth - dimension rectangular window function and a range - dimension rectangular window function to perform range - azimuth two - dimensional truncation; the i - th sub - image x sub_i (t a ,t r ) after truncation, where i = 1, …, N a N r The mathematical expression of / MN is as follows: Among them, the expression of the azimuth dimension rectangular window function is: The expression of the range dimension rectangular window function is: After two-dimensional intercepting by distance and azimuth, (N a N r / MN - 2*N a / M) sub-images with the size of M×(N + 2*N pro ) and 2*N a / M sub-images with the size of M×(N + N pro ) are obtained, denoted as x sub_i (t a , t r ).

4. The interference suppression method based on distance-azimuth two-dimensional intercept processing according to claim 1, wherein, The step of detecting outliers pulse by pulse along the range dimension of the two-dimensional range spectrum includes: The judgment formula for outliers is: wherein, represents the two-dimensional range spectrum of the sub-image at the azimuth sampling time range frequency and the energy value of the sub-image at ; represents the two-dimensional range spectrum X of the sub-image sub_i (t a , f r ) corresponding to all range frequencies at the th azimuth sampling time, median represents the median operation, and c is a constant; When the above equation holds, the is an outlier, that is, a frequency point where interference may exist.

5. The interference suppression method based on distance-azimuth two-dimensional intercept processing according to claim 1, characterized in that The step of constructing a two-dimensional interference mask of the sub-image based on the outliers includes: After performing outlier detection on each azimuth sampling time of the two-dimensional distance spectrum of the sub-image along the range dimension respectively, a two-dimensional interference mask mask of the sub-image is formed. sub_i Among them, the azimuth sampling time range frequency The interference mask value at The calculation formula is: Among them, is the mask indication quantity, indicating the two-dimensional range spectrum corresponding to all range frequencies at the azimuth sampling time ​ 6. The interference suppression method based on distance-azimuth two-dimensional intercept processing according to claim 1, wherein The sub-image range spectrum after interference suppression processing is specifically expressed as: Sub-image distance spectrum X sub_i_after (t a , f r ) is calculated by the formula: X sub_i_after (t a ,f r ) = X sub_i (t a ,f r ).*(1 - mask sub_i ).

7. The interference suppression method based on distance-azimuth two-dimensional truncation processing according to claim 1, wherein The step of constructing the entire SLC image after interference suppression based on the sub-image after interference suppression includes: The sub-image x after interference suppression sub_i_after (t a ,t r ) truncates the protection unit, that is, performs two-dimensional truncation in the range-azimuth direction. For (N a N r / MN - 2*N a / M) sub-images of interference suppression with a size of M×(N + 2*N pro ) and (2*N a / M) sub-images of interference suppression with a size of M×(N + N pro ) respectively truncate the corresponding protection units along the range dimension, obtaining (N a N r / MN) sub-images of interference suppression with a size of M×N, x' sub_i_after (t a ,t r ); then the sub-images are re-stitched in the truncation order to obtain the entire SLC image after interference suppression.

8. A terminal device, comprising a processor, a memory, and a computer program stored in the memory; characterized in that, When the processor executes the computer program, it realizes the interference suppression method based on range-azimuth two-dimensional truncation processing according to any one of claims 1-7.

9. A computer-readable storage medium storing a computer program therein; characterized in that, When the computer program is executed by the processor, it realizes the interference suppression method based on range-azimuth two-dimensional truncation processing according to any one of claims 1-7.

Images