Self-adaptive quantization method and system for ultrahigh-resolution spaceborne SAR (Synthetic Aperture Radar) image

Abstract

The invention discloses an adaptive quantization method and system for an ultrahigh-resolution spaceborne SAR image, and the method comprises the steps: carrying out the statistics of a probability density function of each pixel according to the pixels contained in the SAR image; calculating a cumulative distribution function CDF according to the probability density function of the pixel; performing fine tuning on the cumulative distribution function CDF, so that the fine-tuned cumulative distribution function is strictly and monotonically increased; quantizing the SAR image by taking the fine-tuned cumulative distribution function as a quantization curve to obtain a preliminarily enhanced SAR image; calculating the standard deviation of the original SAR image and the standard deviation of the preliminarily enhanced SAR image so as to determine an adaptive gamma coefficient; and performing gamma transformation on the preliminarily enhanced SAR image by using an adaptive gamma coefficient to realize secondary enhancement of the SAR image. According to the method, the visual effect and readability of the SAR image are improved, and the amplitude and phase information of the SAR image can be recovered with high precision.

Description

Technical Field

[0001] The present invention relates to synthetic aperture radar (SAR) signal processing technology, and particularly to an adaptive quantization method and system for ultra-high resolution spaceborne SAR images. Background Art

[0002] The SAR imaging processing algorithm can convert echo data into ground object scattering coefficients, i.e., SAR images. SAR image quantization is a process of converting a single look complex (SLC) image or amplitude image output by imaging processing software into discrete values according to the quantization bits. To ensure processing accuracy, the SAR imaging processing software stores and processes SAR echo / image data in 32-bit floating-point complex numbers in memory. Considering storage and transmission limitations and adapting to mainstream image protocols, when the SAR imaging processing software outputs standard products, SAR images usually need to be linearly quantized to 16 bits. When the SAR image resolution is at the meter / sub-meter level, the dynamic range of the image pixel brightness is usually within 90 dB, and 16-bit linear quantization can meet the dynamic range requirements. As the image resolution increases, the dynamic range of the SAR image brightness increases sharply. When the image resolution reaches the centimeter level, the image dynamic range can reach 100 dB, exceeding the 16-bit quantization range. If the traditional 16-bit linear quantization method is still used, there will inevitably be quantization losses in the SAR image, affecting the quantitative application of the product. In addition, the SAR image amplitude (pixel brightness) usually conforms to the Rayleigh distribution. In most scenarios (more than 99.9%), the intensity values of pixels are low, while in a very small number of strong scattering targets (such as buildings, ground reflectors, etc.), the intensity of pixels is high. At this time, if the linear quantization method is still used, the utilization rate of quantization bit resources is extremely low, and the overall SAR amplitude image is too dark, and the color level needs to be adjusted to see an image that conforms to the human eye visual characteristics.

[0003] In summary, aiming at the problems of high quantization loss and low quantization resource utilization rate in the quantization of centimeter-level ultra-high resolution SAR images, it is urgent to study adaptive quantization technology to balance data accuracy and storage efficiency, minimize quantization losses as much as possible, output SAR images that conform to the human eye visual characteristics, and provide high-quality products for the quantitative application of backend SAR images.

[0004] Currently, the mainstream method for quantifying spaceborne SAR images is to use simple linear quantization, which evenly divides the image data into different gray levels according to the amplitude size. The advantage of this method is its high computational efficiency and relatively simple implementation, which is suitable for the rapid processing of large-scale data and has high quantization accuracy in SAR images with meter-level resolution. The literature [Gao Wenbin, Jiang Hua, Zhao Cheng, et al. An Improved Adaptive Image Quantization Algorithm [J]. Journal of Beijing Electronic Science and Technology Institute, 2019, 27(02): 28-33.] proposed an improved adaptive image quantization algorithm (abbreviated as the Gao method). This method introduces the concept of the upper limit of the interval to ensure the balance of the image gray distribution. For this purpose, a specific threshold is set, and the pixel points higher than this threshold are processed to weaken their gray values to this threshold. In this way, it can effectively prevent the over-prominence of the area with too high gray values in the image, thereby enhancing the overall visual effect. Subsequently, linear quantization is performed on the processed image data to further optimize the performance of the image and improve its visualization effect. However, this quantization is irreversible, and the quantization loss of strong scattering targets is very large. The logarithmic quantization method optimizes the amplitude distribution characteristics, can effectively utilize the quantization resources, and reduces the quantization loss of weak scattering targets.

[0005] However, due to the Rayleigh distribution characteristics of the SAR image amplitude, linear quantization often results in the overall darkness of the image, especially when the number of strong scattering pixels is small. This darkening phenomenon not only affects the visual effect of the image but also may obscure some important ground object information, thereby reducing the usability of the image. In addition, linear quantization fails to effectively utilize the limited quantization levels, resulting in the waste of some quantization resources. The Gao method may cause significant loss of details by weakening the pixels higher than the threshold, especially in the detail performance of the high-brightness area. This weakening process may make the originally high-brightness part with important visual information become blurred, thus affecting the clarity and readability of the overall image. Although logarithmic quantization theoretically improves the utilization rate of quantization resources, its noise gain is relatively high, which may lead to the enhancement of noise in the image, thereby affecting the interpretation and analysis of the image. Summary of the Invention

[0006] The object of the present invention is to propose an adaptive quantization method and system for ultra-high-resolution spaceborne SAR images.

[0007] The technical solution for realizing the present invention is as follows: An adaptive quantization method for ultra-high-resolution spaceborne SAR images, the steps are as follows:

[0008] Step 1, according to the number of pixels n included in the SAR image, statistically calculate the probability density function P of each pixel n ;

[0009] Step 2, according to the probability density function of the pixels, calculate the cumulative distribution function CDF;

[0010] Step 3, fine-tune the cumulative distribution function CDF so that the fine-tuned cumulative distribution function is strictly monotonically increasing;

[0011] Step 4, use the fine-tuned cumulative distribution function as the quantization curve to quantize the SAR image, and obtain the preliminarily enhanced SAR image;

[0012] Step 5, calculate the standard deviation of the original SAR image and the standard deviation of the preliminarily enhanced SAR image, so as to determine the adaptive gamma coefficient;

[0013] Step 6, perform gamma transformation on the preliminarily enhanced SAR image using the adaptive gamma coefficient to achieve secondary enhancement of the SAR image.

[0014] Furthermore, in Step 1, according to the number of pixels n included in the SAR image, the probability density function P of each pixel is statistically calculated n , which is expressed by the formula:

[0015]

[0016] where N is the total number of pixels in the image.

[0017] Furthermore, in Step 2, according to the probability density function of the pixel, the cumulative distribution function CDF is calculated, which is expressed by the formula:

[0018]

[0019] where K is the maximum pixel value in the original SAR image.

[0020] Furthermore, in Step 3, fine-tune the cumulative distribution function CDF so that the CDF is strictly monotonically increasing, which is expressed by the formula:

[0021] CDF new = CDF + 0.1 * log10(L + 1)

[0022] where CDF new is the fine-tuned cumulative distribution function, and L is the range of the pixel size after quantization of the SAR image, where L ∈ [0, 6555].

[0023] Furthermore, in Step 4, quantize the SAR image according to the fine-tuned cumulative distribution function curve as the quantization curve of the SAR image.

[0024] Y = interp1(edges, CDF new , X, 'line')

[0025] edges = linspace(min(X), max(X), bin)

[0026] where Y is the preliminarily enhanced image, interp1(·) is the interpolation operation, CDF new is the fine-tuned cumulative distribution function, X is the original SAR image, line is linear interpolation, edges is the data interval, linspace(·) is the equal-interval spacing operation. min(·) is the minimum-finding operation, max(·) is the maximum-finding operation, and bin is the specified number of interval points, i.e., the quantization range. In this process, each gray value corresponds to a cumulative distribution function value, and this CDF value represents the cumulative proportion of pixel points less than or equal to this gray level. In this way, the gray values of the original image are re-quantized to new gray levels, thus obtaining the preliminarily enhanced SAR image.

[0027] Further, in step 5, calculate the standard deviation of the original SAR image and the standard deviation of the preliminarily enhanced SAR image, and determine the adaptive gamma coefficient. The formula is expressed as:

[0028]

[0029] where γ new is the designed gamma coefficient, X is the original image, Y is the preliminarily enhanced image, and std(·) is the standard deviation-finding operation.

[0030] Further, in step 6, perform gamma transformation on the preliminarily enhanced SAR image using the adaptive gamma coefficient to achieve secondary enhancement of the SAR image, and determine the quantization curve using the secondarily enhanced SAR image and the original SAR image, and perform SAR image decoding using the inverse function operation. The specific method is as follows:

[0031] Perform gamma transformation on the preliminarily enhanced SAR image using the adaptive gamma coefficient to achieve secondary enhancement of the SAR image. The formula is expressed as:

[0032]

[0033] where I(i, j) is the pixel of the preliminarily enhanced SAR image, I'(i, j) is the pixel of the secondarily enhanced SAR image after gamma transformation, I min is the minimum pixel value of the preliminarily enhanced SAR image, I max is the maximum pixel value of the preliminarily enhanced SAR image, and γ new is the designed gamma coefficient.

[0034] Further, output the secondarily enhanced SAR image as a 16-bit image, and at the same time output the quantization coefficient for the backend quantitative application to restore the original SAR image.

[0035] An adaptive quantization system for ultra-high resolution spaceborne SAR images, implementing the described adaptive quantization method for ultra-high resolution spaceborne SAR images, realizing the adaptive quantization for ultra-high resolution spaceborne SAR images, and respectively executing steps 1 to 7 in six modules.

[0036] A computer device includes a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the described adaptive quantization method for ultra-high resolution spaceborne SAR images is implemented, and the adaptive quantization for ultra-high resolution spaceborne SAR images is realized.

[0037] A computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the described adaptive quantization method for ultra-high resolution spaceborne SAR images is implemented, and the adaptive quantization for ultra-high resolution spaceborne SAR images is realized.

[0038] Compared with the prior art, the significant advantages of the present invention are as follows: 1) The quantization accuracy of the present invention is high, making the output SAR image more in line with the human visual characteristics, and improving the visual effect and readability of the SAR image. 2) The quantization loss of the present invention is small, and the amplitude and phase information of the SAR image can be restored with high precision. Description of the Drawings

[0039] Figure 1 is the gamma curve.

[0040] Figure 2 is the flowchart of the adaptive quantization method for ultra-high resolution spaceborne SAR images.

[0041] Figure 3 are the quantization visualization effects for each scenario, where a - e respectively correspond to airplane, building, forest, ocean, and ship; 1 - 4 respectively correspond to linear quantization, logarithmic quantization, Gao method, and the method of the present invention.

[0042] Figure 4 is the absolute error of the amplitude after decoding of the present invention.

[0043] Figure 5 is the absolute error of the phase after decoding of the present invention.

[0044] Figure 6 are the original data, where (a) is the 512 * 512 original image; (b) is the point target; (c) is the range waveform; (d) is the azimuth waveform.

[0045] Figure 7The data decoded by the method of the present invention, where (a) is a 512*512 original image; (b) is a point target; (c) is a range waveform; (d) is an azimuth waveform. Detailed implementation manners

[0046] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0047] The present invention proposes a method and device for adaptive quantization of ultra-high-resolution spaceborne SAR images. The basic idea is that first, fine-tuning is performed according to the cumulative distribution function of SAR image data, and a quantization curve is constructed based on the fine-tuned cumulative distribution function curve to improve the utilization rate of the quantization space; subsequently, gamma transformation is used to enhance the quantized image, which can make the output image more in line with the human visual characteristics while ensuring the quantization accuracy, improving the visual effect and readability of the image. At the same time, this method has small quantization loss and can accurately restore the amplitude and phase information of the SAR image. The specific implementation steps are as follows:

[0048] Step 1: Statistically calculate the probability density function of all pixels in the SAR image

[0049] According to the number of pixels n included in the SAR image, statistically calculate the probability density function P of each pixel n ;

[0050]

[0051] where N is the total number of pixels in the image.

[0052] The magnitude of pixel n represents the signal intensity of the object reflected back to the radar, and the probability density function can well reflect the brightness change of the SAR image.

[0053] Step 2: Calculate the cumulative distribution function CDF of the SAR image pixels

[0054] Calculate the cumulative distribution function CDF according to the probability density function of the pixels;

[0055]

[0056] where K is the maximum pixel value in the image.

[0057] CDF reflects the distribution of pixel values in the SAR image. Through CDF, the contrast of the image can be adjusted. In the process of SAR image enhancement, using CDF can effectively perform contrast stretching or equalization on the image, making the details in the image more prominent, especially in low-contrast areas.

[0058] Step 3: Fine-tuning of the Cumulative Distribution Function (CDF) of SAR Image Pixels

[0059] Since the Cumulative Distribution Function (CDF) of SAR image pixels calculated in Step 2 is not strictly monotonically increasing, to ensure that each pixel value can be mapped to a unique target pixel value, thus achieving the uniqueness of decoding. Therefore, it is necessary to fine-tune the existing Cumulative Distribution Function (CDF).

[0060] CDF new = CDF + 0.1 * log10(L + 1)

[0061] where CDF new is the fine-tuned Cumulative Distribution Function, and L is the range of the pixel size after quantization of the SAR image, where L ∈ [0, 6555].

[0062] Step 4: Preliminary Enhancement of SAR Image

[0063] Since the Cumulative Distribution Function of SAR image pixels after fine-tuning is a strictly monotonically increasing function, the SAR image is quantized according to the fine-tuned Cumulative Distribution Function curve as the quantization curve of the SAR image.

[0064] Y = interp1(edges, CDF new , X, 'line')

[0065] edges = linspace(min(X), max(X), bin)

[0066] where Y is the preliminarily enhanced image, interp1(·) is the interpolation operation, CDF new is the fine-tuned Cumulative Distribution Function, X is the original SAR image, line is linear interpolation, edges is the data interval, and linspace(·) is the operation of equally dividing the interval. min(·) is the operation of finding the minimum value, max(·) is the operation of finding the maximum value, and bin is the specified number of interval points, that is, the quantization range. In this process, each gray value corresponds to a Cumulative Distribution Function value, and this CDF value represents the cumulative proportion of pixel points less than or equal to this gray level. In this way, the gray values of the original image are re-quantized to new gray levels, thus obtaining the preliminarily enhanced SAR image. In this process, the brightness and contrast of the SAR image will be preliminarily improved, but the strong and weak contrast is not obvious enough.

[0067] Step 5: Design of Adaptive Gamma Coefficient

[0068] Although the gray values of the SAR image have been evenly distributed within the target range, the initially enhanced SAR image performs poorly in terms of strong-weak contrast and fails to fully reveal image details. To address this issue, gamma transformation is introduced as a secondary enhancement step. Gamma transformation is a non-linear image enhancement method. From Figure 1 the gamma curve, it can be clearly seen that when the gamma coefficient is greater than 1, gamma transformation can significantly enhance the contrast in the high-gray region of the image. To design a gamma coefficient that can target SAR images with different characteristics. Therefore, the standard deviation of the original SAR image and the standard deviation of the initially enhanced SAR image are calculated for adaptive gamma coefficient design

[0069]

[0070] where γ new is the designed gamma coefficient, X is the original image, Y is the initially enhanced image, and std(@) is the operation of calculating the standard deviation.

[0071] This adaptive design of the gamma coefficient can effectively enhance the strong-weak contrast of the image according to the characteristics of the SAR image, further improving the visual effect of the image

[0072] Step 6: Secondary enhancement of the SAR image

[0073] According to the gamma coefficient γ designed in Step 5 new perform gamma transformation on the initially enhanced SAR image.

[0074]

[0075] where I(i,j) is the pixel of the initially enhanced SAR image, I'(i,j) is the pixel of the image after gamma transformation, I min is the minimum pixel value of the initially enhanced SAR image, I max is the maximum pixel value of the initially enhanced SAR image, and γ new is the designed gamma coefficient. For the quantized SAR image, the output is a 16-bit image, and the quantization coefficient is also output for the backend quantitative application to restore the original SAR image. The secondary enhancement of the SAR image can effectively balance the adjustment of image brightness and contrast to achieve an ideal visual effect. The technical flow chart of this technology is as Figure 2 shown.

[0076] The present invention also proposes an adaptive quantization system for ultra-high-resolution spaceborne SAR images, which implements the adaptive quantization method for ultra-high-resolution spaceborne SAR images to achieve adaptive quantization of ultra-high-resolution spaceborne SAR images, and six modules respectively execute Steps 1 to 7.

[0077] A computer device includes a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the adaptive quantization method for ultra-high resolution spaceborne SAR images is implemented to achieve adaptive quantization for ultra-high resolution spaceborne SAR images.

[0078] A computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the adaptive quantization method for ultra-high resolution spaceborne SAR images is implemented to achieve adaptive quantization for ultra-high resolution spaceborne SAR images.

[0079] In summary, while ensuring quantization accuracy, the present invention makes the output image more conform to the human visual characteristics, improving the visual effect and readability of the image. At the same time, this method has small quantization loss and can accurately restore the amplitude and phase information of the SAR image. To comprehensively evaluate the effectiveness of this method, the present invention conducts quantization processing experiments on ideal point target simulation data and SAR measured data. The results show that, compared with traditional quantization methods, the present invention has achieved a significant improvement in visual effect. Moreover, it also shows superiority in quantization accuracy, providing new ideas and solutions for high-precision and high-quality quantization processing of SAR images.

[0080] Embodiment

[0081] To verify the effectiveness of the present invention, linear quantization, logarithmic quantization, the Gao method, and the present invention are used for comparative experiments on scenarios such as ocean, forest, building, airplane, and ship. The visualization results after the above linear quantization, logarithmic quantization, the Gao method, and the present invention are finally quantized are as Figure 3 shown.

[0082] According to the original SAR image X and the finally quantized image I'(i, j), the final quantization curve F(·) can be obtained, that is:

[0083] I'(i,j) = F(X)

[0084] According to the inverse function F -1 (·) of the quantization curve F(·) for image decoding:

[0085] X' = F -1 (I'(i,j))

[0086] where X' is the decoded SAR image. This can provide the original SAR image for subsequent quantitative applications to restore.

[0087] Calculate the mean absolute error:

[0088] Noise = mean(|X' - X|)

[0089] Where Noise is the mean absolute error, X' is the decoded SAR image data, X is the original SAR image, |·| is the absolute value operation, and mean(·) is the average value operation. The smaller the mean absolute error, the better the quantization performance. The results are shown in Table 1.

[0090] Calculate the mean square error index:

[0091]

[0092] Where M*N represents the size of the image, X' and X represent the decoded SAR image and the original SAR image respectively, MSE is the mean square error. The smaller the value of MSE, the smaller the error. The results are shown in Table 2.

[0093] Calculate the peak signal-to-noise ratio index:

[0094]

[0095] Where PSNR is the peak signal-to-noise ratio, MSE is the mean square error, L max is the maximum value of the pixels of the quantized SAR image. The higher the PSNR, the better the image quality. The results are shown in Table 3.

[0096] Calculate the standard deviation index of the SAR image, which is generally used to measure the contrast of the image. The larger the standard deviation, the higher the contrast.

[0097]

[0098] Where M and N represent the width and height of the image respectively, I'(i,j) represents the pixel at the i-th row and j-th column of the quantized SAR image, I mean represents the mean value of the pixels of the quantized SAR image. The comparison results are shown in Table 4.

[0099] Calculate the energy index of the image:

[0100] energy = 10log10(∑ M,N I'(i,j) 2 )

[0101] Where energy is the energy value of the SAR image, M and N are the width and height of the image respectively, and I'(i,j) is the size of the pixel of the quantized SAR image. The results are shown in Table 5.

[0102] Calculate the mean absolute phase error index of the decoded SAR image:

[0103]

[0104] where θ is the standard deviation of the decoded phase error, θ1 is the phase of the decoded SAR image, θ2 is the phase of the original SAR image, |·| is the absolute value operation, mean(·) is the average value operation, and the results are shown in Table 6. The distribution histograms of the absolute error of the decoded amplitude and the absolute phase error of the decoded invention are respectively as Figure 3 and Figure 4 shown.

[0105] At the same time, point targets are simulated, and the integral sidelobe ratio (PSLR) and peak sidelobe ratio (ISLR) indexes are calculated for the decoded point target data and the original point target data. The simulation results are as Figure 5 and Figure 6 shown, and the index results are shown in Table 7.

[0106] From Figure 3 it can be seen that the present invention has a good visualization effect. It can be seen from Table 1 to Table 6 that after quantization by the present invention, the SAR image has a small average absolute error, a small mean square error, a high peak signal-to-noise ratio, a high standard deviation, a high energy, and at the same time a small average phase error. From Figure 4 and Figure 5 it can be seen that most of the amplitude errors of the decoded data after quantization by this method are below 10 -5 , and most of the phase errors are below 0.04 degrees. At the same time, from Figure 6 , Figure 7 and Table 7, it can be seen that the quality of the decoded point target image after quantization by this method is almost lossless.

[0107] Table 1 Average absolute error results

[0108]

[0109]

[0110] Table 2 Mean square error results

[0111] Quantization method Linear quantization Logarithmic quantization Gao method The present invention Airplane 6.059E-06 1.087E-09 2.234E-02 7.046E-10 Building 6.057E-06 1.093E-09 2.095E-02 4.311E-10 Forest 6.057E-06 1.063E-09 2.981E-03 2.493E-10 Ocean 4.538E-06 9.072E-10 3.028E-11 1.340E-10 Ship 4.539E-06 8.872E-10 4.744E-03 2.777E-10

[0112] Table 3 Peak signal-to-noise ratio results

[0113] Quantization method Linear quantization Logarithmic quantization Gao method The present invention Airplane 148.5057 185.9661 112.8385 187.8500 Building 148.5066 185.9421 113.1180 189.9835 Forest 148.5066 186.0643 121.5857 192.3622 Ocean 149.7610 186.7524 201.5181 195.0584 Ship 149.7601 186.8492 119.5684 191.8942

[0114] Table 4 Standard deviation results

[0115] Quantization method Linear quantization Logarithmic quantization Gao method The present invention Airplane 3.166E+01 1.525E+03 1.021E+04 1.077E+04 Building 2.585E+01 1.218E+03 7.960E+03 1.151E+04 Forest 1.719E+01 1.055E+03 7.042E+03 1.111E+04 Ocean 8.470E+00 5.833E+02 3.277E+03 7.940E+03 Ship 1.659E+01 7.422E+02 4.405E+03 6.832E+03

[0116] Table 5 Picture energy results

[0117] Quantization method Linear quantization Logarithmic quantization Gao method The present invention Airplane 89.5867 124.4801 140.6442 141.4437 Building 104.6856 140.2126 156.1014 159.2075 Forest 98.9555 136.0242 151.9098 155.4948 Ocean 97.8356 134.9950 149.5908 155.4130 Ship 98.8877 133.7629 148.6839 152.2824

[0118] Average absolute phase error results after decoding of Table 6

[0119] Quantization method Linear quantization Logarithmic quantization Gao method The present invention Airplane 11.8729 0.1677 0.1404 0.0754 Building 9.3387 0.1309 0.0998 0.0518 Forest 8.9707 0.1253 0.0362 0.0473 Ocean 11.6747 0.1772 0.0241 0.0603 Ship 13.8199 0.2124 0.0440 0.0811

[0120] Table 7 Point target simulation results

[0121]

[0122] The technical features of the above embodiments can be combined arbitrarily. For the sake of concise description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in this specification.

[0123] The above-described embodiments merely represent several implementation manners of the present application. The description is relatively specific and detailed, but it should not be construed as a limitation on the scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. An adaptive quantization method for ultra-high resolution spaceborne SAR images, characterized in that: Here are the steps: Step 1: According to the pixels n contained in the SAR image, the probability density function P of each pixel is calculated. n ; Step 2, calculate the cumulative distribution function CDF according to the probability density function of the pixel; Step 3, fine-tuning the cumulative distribution function CDF so that the fine-tuned cumulative distribution function is strictly monotonically increasing; Step 4, using the fine-tuned cumulative distribution function as a quantization curve to quantize the SAR image to obtain a preliminarily enhanced SAR image; Step 5, calculating the standard deviation of the original SAR image and the standard deviation of the SAR image after preliminary enhancement, so as to determine the adaptive gamma coefficient; Step 6: Perform gamma transformation on the initially enhanced SAR image using an adaptive gamma coefficient to achieve secondary enhancement of the SAR image.

2. The adaptive quantization method for ultra-high resolution spaceborne SAR images according to claim 1, characterized in that: Step 2: Calculate the cumulative distribution function CDF based on the probability density function of the pixel. The formula is: Among them, K is the maximum pixel value in the original SAR image.

3. The adaptive quantization method for ultra-high resolution spaceborne SAR images according to claim 1, characterized in that: Step 3: Fine-tune the cumulative distribution function CDF so that it is strictly monotonically increasing. The formula is: CDF new =CDF+0.1*log10(L+1) Among them, CDF new is the fine-tuned cumulative distribution function, L is the range of pixel sizes that need to be quantized for the SAR image, where L∈[0,6555].

4. The adaptive quantization method for ultra-high resolution spaceborne SAR images according to claim 1, characterized in that: Step 4: quantize the SAR image using the fine-tuned cumulative distribution function curve as the quantization curve of the SAR image. The specific method is as follows: Y=interp1(edges,CDF new ,X,'line') edges=linspace(min(X),max(X),bin) Among them, Y is the image after preliminary enhancement, interp1(●) is the interpolation operation, and CDF new is the fine-tuned cumulative distribution function, X is the original SAR image, line is the linear interpolation, edges is the data interval, linspace(·) is the equal interval interval operation, min(●) is the minimum operation, max(●) is the maximum operation, bin is the specified number of interval points, that is, the quantization range, each gray value corresponds to a cumulative distribution function value, and the CDF value represents the cumulative proportion of pixels less than or equal to the gray level. In this way, the gray value of the original image is requantized to a new gray level, thereby obtaining a preliminarily enhanced SAR image.

5. The adaptive quantization method for ultra-high resolution spaceborne SAR images according to claim 1, characterized in that: Step 5, calculate the standard deviation of the original SAR image and the standard deviation of the SAR image after preliminary enhancement, and determine the adaptive gamma coefficient, which is expressed as follows: Among them, γ new is the designed gamma coefficient, X is the original image, Y is the image after preliminary enhancement, and std(●) is the standard deviation operation.

6. The adaptive quantization method for ultra-high resolution spaceborne SAR images according to claim 1, characterized in that: Step 6, using the adaptive gamma coefficient to perform gamma transformation on the initially enhanced SAR image to achieve secondary enhancement of the SAR image, and using the secondary enhanced SAR image and the original SAR image to determine the quantization curve, and perform inverse function operation to decode the SAR image. The specific method is: The adaptive gamma coefficient is used to perform gamma transformation on the initially enhanced SAR image to achieve secondary enhancement of the SAR image. The formula is expressed as: Where I(i,j) is the pixel of the primary enhanced SAR image, I'(i,j) is the pixel of the secondary enhanced SAR image after gamma transformation, and I min is the minimum pixel value of the preliminary enhanced SAR image, I max is the maximum pixel value of the preliminary enhanced SAR image, γ new is the designed gamma coefficient.

7. The adaptive quantization method for ultra-high resolution spaceborne SAR images according to claim 1, characterized in that: The secondary enhanced SAR image is output as a 16-bit image, and the quantization coefficient is output at the same time for the back-end quantitative application to restore the original SAR image.

8. An adaptive quantization system for ultra-high resolution spaceborne SAR images, implementing the adaptive quantization method for ultra-high resolution spaceborne SAR images described in any one of claims 1 to 6, realizing adaptive quantization for ultra-high resolution spaceborne SAR images, and executing steps 1 to 7 in six modules respectively.

9. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the adaptive quantization method for ultra-high resolution spaceborne SAR images according to any one of claims 1 to 7 is implemented to realize adaptive quantization for ultra-high resolution spaceborne SAR images.

10. A computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the method for adaptive quantization of ultra-high resolution spaceborne SAR images according to any one of claims 1 to 7 is implemented to achieve adaptive quantization of ultra-high resolution spaceborne SAR images.

Images