Satellite-borne SAR scene matching curve imaging azimuth fuzzy detection and suppression method

Abstract

The invention discloses a satellite-borne SAR (synthetic aperture radar) scene matching curve imaging azimuth ambiguity detection and suppression method, which comprises the following steps of: 1, modeling azimuth ambiguity, and respectively giving an azimuth ambiguity echo analytic expression, range migration correction and azimuth phase compensation items; 2, refocusing n-order azimuth ambiguity according to range migration correction and an azimuth phase compensation item; 3, the entropy difference of the SAR image and the n-order fuzzy refocusing image is calculated in a blocking mode, and fuzzy region detection is completed through a CFAR algorithm; 4, performing filtering processing by using a Wiener filter to generate a low-blur and low-resolution image; and in combination with the detection result of the azimuth blurred region, a low-blurring and high-resolution image is obtained by replacing sub-blocks obtained by detection. According to the method, the problem of orientation blur in a satellite-borne SAR scene matching curve imaging mode can be solved, and the resolution of an original image is reserved while the image quality is effectively improved.

Description

Technical Field

[0001] The invention belongs to the technical field of synthetic aperture radar (SAR for short), and in particular relates to a method for detecting and suppressing azimuth fuzziness of spaceborne SAR scene matching curve imaging. Background Art

[0002] Azimuth ambiguity is an important indicator for the quality assessment of spaceborne SAR images. The imaging band of the traditional spaceborne SAR imaging mode is along the direction of the satellite orbit, and the geographical direction is single; the imaging band of the spaceborne SAR scene matching curve imaging mode may not be parallel to the satellite orbit direction, and the geographical direction is complex and changeable. Since the spaceborne SAR antenna pattern has sidelobe energy in addition to the energy required for the observation area, the echo power observed in the sidelobe area enters the mainlobe observation area to form fuzzy energy, and the received undesired fuzzy energy and the main image area energy are mixed in the image, affecting the high-resolution SAR imaging quality and the effectiveness of ground observation. In the spaceborne SAR scene matching curve imaging, the scene direction, echo delay, and Doppler history change dramatically with azimuth time. When the azimuth ambiguity is relatively poor, the antenna sidelobe energy will cause great interference to the SAR image, resulting in false targets and serious reduction of image quality, which greatly limits the further interpretation and application of spaceborne SAR images. In short, suppressing the azimuth ambiguity of SAR images in the scene matching curve imaging mode can improve the quality and interpretability of SAR images.

[0003] Conventional azimuth ambiguity suppression methods are mostly performed through post-signal processing for traditional spaceborne SAR imaging modes, such as stripe mode, which has poor adaptability to scene configuration and Doppler process spatial variations; at the same time, azimuth ambiguity suppression through signal processing alone is often accompanied by resolution loss, resulting in further deterioration of image quality. Therefore, it is necessary to deeply analyze the azimuth ambiguity characteristics in the non-along-track scene matching curve imaging mode of spaceborne SAR, study the azimuth ambiguity formation mechanism, and carry out research on low-resolution loss suppression signal processing algorithms for azimuth ambiguity problems. Summary of the invention

[0004] In view of this, the present invention provides an azimuth ambiguity suppression algorithm with low resolution loss in spaceborne SAR scene matching curve imaging, which can solve the azimuth ambiguity problem in the spaceborne SAR scene matching curve imaging mode, effectively improve the SAR image quality while retaining the resolution of the original image.

[0005] The technical solution for implementing the present invention is as follows:

[0006] The method for detecting and suppressing azimuth blur in spaceborne SAR scene matching curve imaging comprises the following steps:

[0007] Step 1: Model the azimuth ambiguity, and give the azimuth ambiguity echo analytical expression, range migration correction compensation term and azimuth phase compensation term respectively;

[0008] Step 2: refocusing the n-order azimuth blur based on the SAR image according to the range migration correction compensation term and the azimuth phase compensation term;

[0009] Step 3: Calculate the image entropy difference between the SAR image and the n-th order azimuthally blurred refocused image, and use the CFAR algorithm to complete the azimuthally blurred area detection;

[0010] Step 4: Use the Wiener filter to filter the SAR image to generate a low-blur, low-resolution image; combined with the azimuth blur area detection results, the low-blur, high-resolution image is obtained by replacing the detected sub-blocks to complete blur suppression.

[0011] Furthermore, the range migration correction compensation term and the azimuth phase compensation term in step 2 are respectively:

[0012]

[0013] Where R is the reference slope distance, R am is the reference slope distance of the fuzzy zone, f η is the azimuth frequency, ΔR(f η ) is the distance migration correction compensation item, is the azimuth phase compensation term.

[0014] Furthermore, the image entropy in step 3 is:

[0015]

[0016] Among them, I(x, y) is the gray value of the image, S is the number of times different gray levels appear in the image, and M and N are the total number of sample points in the azimuth and distance directions of the image.

[0017] Furthermore, in step 3, the CFAR detection method is used to detect the entropy difference change before and after refocusing to achieve the detection of the main blurred area target:

[0018] ΔE=E refoused -E original (3)

[0019] Among them, E refoused represents the image entropy after refocusing, E original represents the initial SAR image entropy, and ΔE represents the entropy difference before and after processing.

[0020] Furthermore, the Wiener filter in step 4 is:

[0021]

[0022] Among them, H(f η ) is the azimuth frequency domain filter, f η is the azimuth frequency, σ and σ0 are the backscattering coefficients of the main image target and the blurred target respectively, σ t is the white hot noise power density, and G(·) is the antenna gain.

[0023] Beneficial effects:

[0024] 1. The present invention solves the azimuth ambiguity problem in non-along-track scene matching curve imaging of spaceborne SAR, and can effectively improve the SAR image quality and suppress azimuth ambiguity.

[0025] 2. Steps 1 and 2 of the present invention provide the analytical expression of the azimuth ambiguous echo, the range migration correction compensation term and the azimuth phase compensation term of the mode, which solves the problem of focusing on the azimuth ambiguous echo and lays a research foundation for subsequent detection suppression.

[0026] 3. Step 3 of the present invention proposes a blur region detection algorithm based on image entropy difference, which can complete blur region detection in the case of azimuth blur space variation;

[0027] 4. Step 4 of the present invention proposes a blur suppression algorithm based on low-blur image replacement. After processing using the Wiener filtering method, a low-blur, low-resolution image can be obtained. The azimuth blur detection result is used to replace the low-blur image to obtain an overall high-resolution, low-blur SAR image, thereby achieving azimuth blur suppression with low resolution loss. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 It is a flow chart of the method for detecting and suppressing azimuth blur in spaceborne SAR scene matching curve imaging according to the present invention;

[0029] Figure 2 Schematic diagram of the causes of azimuth ambiguity in non-along-track scene matching curve imaging of spaceborne SAR according to the present invention; (a) geometric configuration of azimuth ambiguity of spaceborne SAR, (b) aliasing of azimuth ambiguity energy in Doppler domain;

[0030] Figure 3 It is a schematic diagram of the formation of azimuth blur in the conventional range-Doppler domain imaging algorithm;

[0031] Figure 4 Schematic diagram of azimuth fuzzy focusing imaging according to the method of the present invention;

[0032] Figure 5Schematic diagram of ±1-order azimuth fuzzy focusing imaging according to the method of the present invention; (a) schematic diagram of +1-order azimuth fuzzy range migration correction, (b) schematic diagram of -1-order azimuth fuzzy range migration correction, (c) +1-order azimuth fuzzy focusing imaging, (d) -1-order azimuth fuzzy focusing imaging, (e) +1-order azimuth fuzzy focusing imaging azimuth peak sidelobe ratio, (f) -1-order azimuth fuzzy focusing imaging azimuth peak sidelobe ratio;

[0033] Figure 6 Schematic diagram of the method of the present invention after suppressing azimuth ambiguity in an embodiment; (a) SAR image before suppressing azimuth ambiguity, (b) azimuth peak sidelobe ratio before suppressing azimuth ambiguity, (c) SAR image after suppressing azimuth ambiguity, (d) azimuth peak sidelobe ratio after suppressing azimuth ambiguity. DETAILED DESCRIPTION

[0034] The present invention is described in detail below with reference to the accompanying drawings and embodiments.

[0035] The flowchart of the method for detecting and suppressing azimuth blur in spaceborne SAR scene matching curve imaging according to the present invention is as follows: Figure 1 As shown, the present invention comprises the following steps:

[0036] Step 1: Model the azimuth ambiguity and give analytical expressions of azimuth ambiguity echoes respectively;

[0037] In spaceborne SAR, the transmission and reception of the same echo are separated by several PRTs. Figure 2 Taking the first fuzzy area in (a) as an example, the radar receiver receives the echo from the fuzzy area and the echo from the main imaging area at the same time. During the movement of the satellite, the three areas will form a continuous Doppler frequency, resulting in exceeding the system sampling bandwidth. Since the system performs azimuth discrete sampling according to the PRF in azimuth, the signal will be aliased within the sampling bandwidth. The aliased signal is focused by the imaging processor and eventually forms azimuth ambiguity in the SAR image. When not affected by the weighting in the azimuth plane, the one-way radiation pattern of the satellite-borne SAR antenna beam can be approximated as a sinc function related to the angle θ:

[0038]

[0039] In the formula, G(·) represents the echo azimuth gain, L a represents the azimuth dimension of the antenna, λ represents the wavelength of the radar carrier frequency, and θ represents the incident angle.

[0040] The demodulated signal of SAR point target reception can be written as:

[0041]

[0042] The slant distance R(η) at different azimuths is:

[0043]

[0044] Where τ and η are the azimuth time and distance time respectively, σ0 is the backscattering coefficient of the point target, and η c is the beam center crossing time, V r is the equivalent radar speed, R0 is the shortest slant range, c is the speed of light, K r is the range modulation frequency, and w(·) represents the rectangular envelope.

[0045] In the design of spaceborne SAR systems, the PRF is generally 1.1-1.2 times the Doppler bandwidth. Limited azimuth sampling samples the theoretically infinite Doppler spectrum, resulting in aliasing of the Doppler spectrum, which is ultimately reflected as azimuth blur in the SAR image. Figure 2 As shown in (b), when the Doppler frequency of the echo in this area is [f d -PRF / 2,f d +PRF / 2], this area is called the main imaging area. d -PRF / 2±n·PRF,f d +PRF / 2±n·PRF], then this area is called the nth (n≠0, n∈Z) fuzzy area, where f d is the Doppler center frequency. The SAR point target azimuth ambiguity signal can be written as:

[0046]

[0047] The slant range R corresponding to the fuzzy signal at different azimuths am (η) is:

[0048]

[0049] Where PRT is the reciprocal of the pulse repetition frequency PRF.

[0050] Affected by the echo delay, the range migration of the echo signal after one-dimensional range pulse compression shows a parabolic characteristic. After range migration correction, the migration line of the main imaging area is completely straightened, while the migration line of the blurred area aliased into the main imaging area is tilted. The tilted migration line will cause the imaging result to have range dispersion and azimuth defocus after azimuth compression, such as Figure 3 shown.

[0051] Transforming the SAR point target signal into the range Doppler domain, we have:

[0052]

[0053] The range migration equation in the range Doppler domain is R rd (f η)for:

[0054]

[0055] According to the azimuth ambiguity characteristics, the distance migration equation R rd (f η ) is rewritten as:

[0056]

[0057] The n-order azimuth ambiguity echo signal in the range Doppler domain is:

[0058]

[0059] By using the precise expression of the n-order azimuth ambiguity echo signal, the azimuth ambiguity can be focused in the image domain, such as Figure 4 shown.

[0060] Step 2: refocusing the n-order azimuth blur based on the SAR image according to the range migration correction compensation term and the azimuth phase compensation term;

[0061] The range migration correction compensation term and the azimuth phase compensation term are:

[0062]

[0063] According to equations (8) and (9), it can be further written as:

[0064]

[0065] After performing azimuth IFFT on the image, range motion correction compensation and azimuth phase compensation are performed, and finally azimuth FFT is performed to obtain the blurred focused image after refocusing.

[0066] Step 3: Calculate the image entropy difference between the SAR image and the n-th order azimuthally blurred refocused image, and use the CFAR algorithm to complete the azimuthally blurred area detection;

[0067] Through the precise expression of the n-order azimuth ambiguity echo signal, a derivative image can be generated based on the original SLC image, including the focused imaging results of each ambiguity area to be suppressed. In order to accurately locate the azimuth ambiguity distribution area, we need to find the azimuth ambiguity distribution area through focusing accuracy. In general, the focusing quality of SAR images can be measured by image entropy:

[0068]

[0069] Where I(x,y) is the grayscale value of the image, M and N are the total number of sample points in the azimuth and distance directions. According to the definition of entropy, the smaller the image entropy, the better the image focusing effect.

[0070] The azimuthally blurred focus imaging results include not only the azimuthally blurred targets, but also the defocus results of the main imaging area and other order blurred areas, whose backscattering intensity is weaker than that of the blurred focus area. The problem of detecting targets in the blurred focus area against the background of the defocused main imaging area can be transformed into the problem of detecting entropy values ​​after image segmentation. The detection of targets in the main blurred area can be achieved by using the CFAR detection method on the entropy difference change before and after refocusing:

[0071] ΔE=E refoused -E original (17)

[0072] Among them, E refoused represents the image entropy after refocusing, E original represents the initial SAR image entropy.

[0073] The algorithm detects by comparing the entropy difference between the blurred focused image and the SAR initial image, and uses the entropy value of the image in the target window and the false alarm rate (PFA) and the entropy threshold obtained by background clutter image entropy modeling to achieve adaptive target detection. The background clutter is assumed to be a Gaussian distribution Φ, with mean and variance μ and σ respectively. The threshold T can be expressed as:

[0074]

[0075] The essence of azimuthally blurred target detection is to select a suitable threshold T so that the image meets the false alarm rate P. a It can be divided into two distinguishable regions:

[0076]

[0077] Step 4: Use the Wiener filtering method to generate a low-blur, low-resolution image; combine the azimuth blur area detection results, and obtain a low-blur, high-resolution image by replacing the detection sub-block to complete the azimuth blur suppression.

[0078] Since the blur energy is often concentrated at both ends of the Doppler spectrum, the azimuth blur energy can be reduced by azimuth spectrum weighting by taking advantage of the low correlation of the blur spectrum, or the area with large blur energy can be screened out by the difference in energy distribution of the image before and after spectrum weighting. However, this method also changes the spectrum shape of the target in the main area, which will interfere with the imaging result of the real target. The most direct impact is that the resolution of the area covered by the blur is greatly reduced. Therefore, an azimuth blur suppression algorithm based on low blur image replacement is proposed. By using the azimuth blur detection results and the low blur SAR image with spectrum truncation, the fuzzy area replacement is used to achieve the suppression of high-resolution, high-blur SAR images into high-resolution, low-blur high-quality SAR images.

[0079] SAR echo energy can be expressed in the range Doppler domain as:

[0080] s rd (f η ,τ)=Γ0G0(f η ,τ)+Γ + G + (f η ,τ)+Γ_G_(f η ,τ) (20)

[0081] Where Γ represents the backscattering coefficient after Fourier transformation.

[0082] At the same time, the unambiguous signal and the echo signal affected by additive white hot noise can be defined as:

[0083] s0(f η ,τ)=Γ0G0(f η ,τ) (21)

[0084] X(f η ,τ)=Γ0G0(f η ,τ)+Γ + G + (f η ,τ)+Γ_G_(f η ,τ)+T(f η ) (twenty two)

[0085] The Wiener filter can filter out the azimuth ambiguity by using the azimuth samples of the image, minimizing the mean square error between the "true" unambiguous signal and the filtered signal. The power spectral density of the main image target and the blurred target is as follows:

[0086]

[0087] where E[·] represents the mean and σ represents the power spectral density of the backscatter coefficient, which is assumed to be statistically stable within the spatial filter range.

[0088] Based on the assumptions, the expression of the Wiener filter is:

[0089]

[0090] Considering the above expression, assuming that the echo data and the unambiguous data are both zero-mean and uncorrelated, we can obtain from equation (24):

[0091]

[0092] where σ t is the white hot noise power density.

[0093] Considering that σ=σ + ≈σ -, formula (25) can be further rewritten as:

[0094]

[0095] After filtering (26), a low-blur, low-resolution image with azimuth blur energy removed can be obtained, and the azimuth blur area is replaced by the azimuth blur detection result. Thanks to the Wiener filtering method and the local image processing idea, azimuth blur suppression with almost no resolution loss is finally achieved.

[0096] Simulation experiment: The simulation parameters of the azimuth blur detection and suppression method for spaceborne SAR scene matching curve imaging are shown in Table 1.

[0097] Table 1 List of simulation parameters for the azimuth blur detection and suppression method of spaceborne SAR scene matching curve imaging

[0098]

[0099] In order to verify the advantages of the azimuth ambiguity detection and suppression method for spaceborne SAR scene matching curve imaging in solving the azimuth ambiguity problem in the non-along-track scene matching curve imaging mode of spaceborne SAR, the azimuth ambiguity suppression method of the present invention is used to design the azimuth ambiguity suppression of point targets under the parameters in Table 1.

[0100] exist Figure 5 The focusing imaging results of the proposed method after range migration correction compensation and azimuth phase compensation are given in FIG. Figure 5 (a) and (b) show that after the range migration correction compensation, the range migration line in the ±1-order fuzzy area is corrected to a straight line; Figure 5 (c) and (d) are the focused images of the ±1-order blurred area after azimuth phase compensation. It is obvious that the method of the present invention has a good focusing effect on the azimuth blurred echo. Figure 5 (e) and (f) respectively give Figure 5 The azimuth peak sidelobe ratio imaging performance indicators of (c) and (d) show that the focusing effect is good as reflected by the peak amplitude. Figure 6 (a) and (b) show the focusing result of the main target image and the azimuth peak sidelobe ratio before the processing of the method of the present invention. It can be seen that the azimuth peak sidelobe before the processing is -37.35dB; Figure 6 (c) and (d) show the focusing result of the main target image processed by the method of the present invention and the azimuth peak sidelobe ratio. It can be seen that the azimuth peak sidelobe before processing is -54.78dB, and the azimuth peak sidelobe after processing is reduced by 17.43dB, which significantly reduces the influence of azimuth blur on the SAR image quality while retaining the resolution of the original image. The suppression effect is shown in Table 2.

[0101] Table 2. Suppression effect of the azimuth ambiguity suppression method proposed in this paper

[0102] Azimuth Peak Sidelobe Ratio Azimuth resolution Before inhibition -37.35dB 1.51m After inhibition -54.78dB 1.51m

[0103] In summary, the above are only preferred embodiments of the present invention and are not intended to limit the protection scope of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for detecting and suppressing azimuth blur in spaceborne SAR scene matching curve imaging, characterized in that: The following steps are involved: Step 1: Model the azimuth ambiguity, and give the azimuth ambiguity echo analytical expression, range migration correction compensation term and azimuth phase compensation term respectively; Step 2: refocusing the n-order azimuth blur based on the SAR image according to the range migration correction compensation term and the azimuth phase compensation term; Step 3: Calculate the image entropy difference between the SAR image and the n-th order azimuthally blurred refocused image, and use the CFAR algorithm to complete the azimuthally blurred area detection; Step 4: Use the Wiener filter to filter the SAR image to generate a low-blur, low-resolution image; combined with the azimuth blur area detection results, a low-blur, high-resolution image is obtained by replacing the detection sub-block to complete blur suppression.

2. The method for detecting and suppressing azimuth blur in spaceborne SAR scene matching curve imaging according to claim 1, characterized in that: The range migration correction compensation term and the azimuth phase compensation term in step 2 are respectively: Where R is the reference slope distance, R am is the reference slope distance of the fuzzy zone, f η is the azimuth frequency, ΔR(f η ) is the distance migration correction compensation item, is the azimuth phase compensation term.

3. The method for detecting and suppressing azimuth blur in spaceborne SAR scene matching curve imaging according to claim 1, characterized in that: The image entropy in step 3 is: Among them, I(x, y) is the gray value of the image, S is the number of times different gray levels appear in the image, and M and N are the total number of sample points in the azimuth and distance directions of the image.

4. The method for detecting and suppressing azimuth blur in spaceborne SAR scene matching curve imaging according to claim 1, characterized in that: In step 3, the CFAR detection method is used to detect the entropy difference before and after refocusing to detect the main blurred area target: ΔE=E refoused -E original Among them, E refoused represents the image entropy after refocusing, E original represents the initial SAR image entropy, and ΔE represents the entropy difference before and after processing.

5. The method for detecting and suppressing azimuth blur in spaceborne SAR scene matching curve imaging according to claim 1, characterized in that: The Wiener filter in step 4 is: Among them, H(f η ) is the azimuth frequency domain filter, f η is the azimuth frequency, σ and σ0 are the backscattering coefficients of the main image target and the blurred target respectively, σ t is the white hot noise power density, and G(·) is the antenna gain.

Images