Method and system for rapidly intercepting ISAR image target
By calculating the grayscale integral projection vector of ISAR images and using a dynamic threshold mechanism, the high complexity and adaptability of ISAR image target extraction methods in low signal-to-noise ratio environments are solved. This achieves fast and accurate target extraction and noise suppression, and is applicable to various target types such as ships and aircraft.
Patent Information
- Application Number
- CN202511526192.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-10-24
- Publication Date
- 2026-01-13
AI Technical Summary
Existing ISAR image target extraction methods are computationally complex or labor-intensive in low signal-to-noise ratio environments, making it difficult to effectively handle complex noise and multiple target types.
By receiving ISAR images and calculating grayscale integral projection vectors in two directions, threshold coefficients and grayscale thresholds are calculated using signal-to-noise ratios. Pixels that do not meet the thresholds are removed, the coordinates of the target image are determined, integral projection is used to suppress noise, and the threshold is dynamically adjusted to adapt to different target types.
It enables fast and accurate target image capture in low signal-to-noise ratio environments, improves processing speed and noise resistance, adapts to various target types, and achieves millisecond-level response on embedded platforms.
Smart Images
Figure CN121330004A_ABST
Abstract
Description
Technical Field
[0001] This application belongs to the field of radar image processing technology, and specifically relates to a method and system for rapid target acquisition in ISAR images. Background Technology
[0002] Currently, real-time detection of targets such as ships in low signal-to-noise ratio environments requires inverse synthetic aperture radar (ISAR) to extract target regions from images. There are various extraction methods, such as fast extraction techniques based on threshold segmentation, intelligent extraction based on shape analysis, and deep learning-based methods.
[0003] While threshold-based segmentation methods have low computational complexity, they are sensitive to complex noise and are only suitable for ship targets in low-noise backgrounds. Intelligent interception methods based on shape analysis require extracting the outline of the ship target first, resulting in high computational complexity. Deep learning-based methods require building an ISAR ship target dataset in advance and enhancing the model's capabilities through data, which involves a significant workload.
[0004] Therefore, how to more effectively crop the target region in an image is a problem that needs to be solved. Summary of the Invention
[0005] The purpose of this application is to provide a method and system for fast target capture in ISAR images, so as to solve the problems of high complexity or large workload of existing capture methods.
[0006] The technical solution of this application is: a method for fast target capture in ISAR images, comprising:
[0007] Receive ISAR images, sum the gray values of the ISAR images in two directions, and generate integral projection vectors in two directions;
[0008] Calculate the maximum and average gray values in the integral projection vectors of two directions, calculate the signal-to-noise ratio using the maximum and average gray values, and calculate the threshold coefficient based on the signal-to-noise ratio.
[0009] The grayscale threshold value is calculated based on the threshold coefficient and the average grayscale value.
[0010] The gray value of each pixel in the integral projection vector is compared with the gray value threshold. Pixels that do not meet the gray value threshold are removed to obtain the pixels and distance coordinates of the target image.
[0011] The gray values of the pixels in the target image are summed along the distance direction to obtain the horizontal and vertical coordinates of the target image.
[0012] Preferably, the signal-to-noise ratio is calculated as follows:
[0013] Calculate the ratio of the maximum grayscale value to the average grayscale value, and determine if it is less than 9. If it is, set the signal-to-noise ratio to 1.5; otherwise, set the signal-to-noise ratio to 1.
[0014] Preferably, the grayscale threshold value is calculated by multiplying the signal-to-noise ratio by the average grayscale value to obtain the grayscale threshold value.
[0015] Preferably, distance values in two directions are set respectively, and the coordinate values of discrete points that exceed the grayscale threshold are statistically analyzed. Then, the coordinate values of the discrete points are compared with the distance values in the two directions respectively. If the distance values in the two directions are satisfied, the discrete point is determined to belong to the target image; otherwise, the discrete point is deleted.
[0016] As one specific implementation, an ISAR image target rapid capture system includes:
[0017] The image preprocessing module is used to receive ISAR images, sum the gray values in the azimuth direction of the ISAR images, and generate an integral projection vector.
[0018] The signal-to-noise ratio calculation module is used to calculate the maximum and average gray values in the integral projection vector, calculate the signal-to-noise ratio using the maximum and average gray values, and calculate the threshold coefficient based on the signal-to-noise ratio.
[0019] The grayscale threshold calculation module is used to calculate the grayscale threshold value based on the threshold coefficient and the average grayscale value.
[0020] The target image determination module is used to compare the gray value of each pixel in the integral projection vector with the gray value threshold, and remove pixels that do not meet the gray value threshold to obtain the pixels and distance coordinates of the target image.
[0021] The coordinate determination module is used to sum the gray values of the pixels in the target image along the distance direction to obtain the horizontal and vertical coordinates of the target image.
[0022] Preferably, in the signal-to-noise ratio calculation module, the ratio of the maximum value to the average value of the grayscale value is calculated, and it is determined whether it is less than 9. If it is, the signal-to-noise ratio is set to 1.5; if not, the signal-to-noise ratio is set to 1.
[0023] Preferably, in the grayscale threshold calculation module, the signal-to-noise ratio is multiplied by the average grayscale value to obtain the grayscale threshold value.
[0024] Preferably, in the grayscale threshold calculation module, distance values in two directions are set respectively. The coordinate values of discrete points that exceed the grayscale threshold are statistically analyzed. Then, the coordinate values of the discrete points are compared with the distance values in the two directions respectively. When the distance values in the two directions are satisfied, the discrete point is determined to belong to the target image; otherwise, the discrete point is deleted.
[0025] The method and system for rapid target capture in ISAR images disclosed in this application have the following advantages:
[0026] Improved real-time performance: It avoids iterative calculations, resulting in faster processing speeds compared to traditional methods.
[0027] Strong noise resistance: It suppresses diffuse noise through integral projection, improving the completeness of target capture.
[0028] Adaptive scenarios: The dynamic threshold mechanism can adapt to various target types such as ships and aircraft, and achieve millisecond-level response on embedded platforms (such as TI C6678). Attached Figure Description
[0029] To more clearly illustrate the technical solutions provided in this application, the accompanying drawings will be briefly described below. Obviously, the drawings described below are merely some embodiments of this application.
[0030] Figure 1 This is a flowchart of the threshold coefficient calculation method in this application;
[0031] Figure 2 This is a flowchart of the method for calculating the target area in this application;
[0032] Figure 3 This is a schematic diagram of an ISAR image from a specific example of this application;
[0033] Figure 4 A schematic diagram comparing the point and azimuth integral projection curves that cross the threshold in this application;
[0034] Figure 5 This is a schematic diagram comparing the integral projection curves of points that cross the threshold and the distances in this application. Detailed Implementation
[0035] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0036] A fast target extraction method for ISAR images is proposed, which can provide target material for target recognition, target size estimation and other functions.
[0037] Includes the following steps:
[0038] Step S100, Data Input: Input an ISAR image, which is a grayscale image with a value range of 0 to 255 and dimensions of azimuth × range.
[0039] Receive ISAR images, sum the gray values in two directions to generate integral projection vectors in two directions.
[0040] This operation compresses a two-dimensional image into a one-dimensional energy distribution curve, enhancing the energy characteristics of the target scattering points while suppressing scattered noise (such as stripe noise and dot noise).
[0041] Step S200: Calculate the maximum and average gray values in the integral projection vectors of the two directions; calculate the signal-to-noise ratio using the maximum and average gray values; and calculate the threshold coefficient based on the signal-to-noise ratio. Figure 1 , .
[0042] Preferably, the signal-to-noise ratio is calculated as follows:
[0043] Calculate the ratio of the maximum grayscale value to the average grayscale value, and determine if it is less than 9. If it is, set the signal-to-noise ratio to 1.5; otherwise, set the signal-to-noise ratio to 1.
[0044] Specifically, such as Figure 1 As shown:
[0045] Here, `max` calculates the maximum value in the array, and `mean` calculates the mean of the array. This logic determines the scene noise level through extreme value distribution: at low SNR, the noise suppression capability is improved by increasing the threshold coefficient (1.5 times the mean); at high SNR, the mean is used directly as the coefficient to avoid over-segmentation of the target.
[0046] Step S300: Calculate the grayscale threshold value based on the threshold coefficient and the average grayscale value.
[0047] Preferably, the grayscale threshold value is calculated by multiplying the signal-to-noise ratio by the average grayscale value to obtain the grayscale threshold value.
[0048] Based on the Rayleigh distribution characteristics of background noise, the grayscale threshold value is calculated as follows: threshold = ThreCoe * mean(temp), where ThreCoe is the signal-to-noise ratio and mean(temp) is the average grayscale value.
[0049] The Rayleigh distribution model assumes that the background noise follows a Gaussian distribution and effectively separates noise from the target scattering point by dynamically adjusting the threshold value based on the mean.
[0050] Step S400: Compare the gray value of each pixel in the integral projection vector with a gray value threshold, and discard pixels that do not meet the gray value threshold to obtain the pixel points and distance coordinates of the target image, such as... Figure 2 .
[0051] Preferably, the specific method for obtaining the pixel points and distance coordinates of the target image is as follows: set distance values in two directions respectively, count the coordinate values of discrete points that exceed the grayscale threshold, and then compare the coordinate values of the discrete points with the distance values in the two directions respectively. When the distance values in the two directions are satisfied, it is determined that the discrete point belongs to the target image; otherwise, the discrete point is deleted.
[0052] Step S500: Sum the gray values of the pixels in the target image along the distance direction to obtain the horizontal and vertical coordinates of the target image.
[0053] In summary, this application has the following advantages:
[0054] Improved real-time performance: It avoids iterative calculations, resulting in faster processing speeds compared to traditional methods.
[0055] Strong noise resistance: It suppresses diffuse noise through integral projection, improving the completeness of target capture.
[0056] Adaptive scenarios: The dynamic threshold mechanism can adapt to various target types such as ships and aircraft, and achieve millisecond-level response on embedded platforms (such as TI C6678).
[0057] As another specific implementation, an ISAR image target rapid capture system includes:
[0058] The image preprocessing module is used to receive ISAR images, sum the gray values in the azimuth direction of the ISAR images, and generate an integral projection vector.
[0059] The signal-to-noise ratio calculation module is used to calculate the maximum and average gray values in the integral projection vector, calculate the signal-to-noise ratio using the maximum and average gray values, and calculate the threshold coefficient based on the signal-to-noise ratio.
[0060] The grayscale threshold calculation module is used to calculate the grayscale threshold value based on the threshold coefficient and the average grayscale value.
[0061] The target image determination module is used to compare the gray value of each pixel in the integral projection vector with the gray value threshold, and remove pixels that do not meet the gray value threshold to obtain the pixels and distance coordinates of the target image.
[0062] The coordinate determination module is used to sum the gray values of the pixels in the target image along the distance direction to obtain the horizontal and vertical coordinates of the target image.
[0063] Preferably, in the signal-to-noise ratio calculation module, the ratio of the maximum value to the average value of the grayscale value is calculated, and it is determined whether it is less than 9. If it is, the signal-to-noise ratio is set to 1.5; if not, the signal-to-noise ratio is set to 1.
[0064] Preferably, in the grayscale threshold calculation module, the signal-to-noise ratio is multiplied by the average grayscale value to obtain the grayscale threshold value.
[0065] Preferably, in the grayscale threshold calculation module, distance values in two directions are set respectively. The coordinate values of discrete points exceeding the grayscale threshold are statistically analyzed, and then the coordinate values of the discrete points are compared with the distance values in both directions. If the distance values in both directions are satisfied, the discrete point is determined to belong to the target image; otherwise, the discrete point is deleted. This is used to distinguish discrete pixels with higher brightness.
[0066] The following is an example to illustrate this:
[0067] 1. Input data: ISAR image, such as Figure 3 As shown.
[0068] Azimuth integral projection calculation: After summing along the azimuth direction, the energy projection vector temp and the peak signal-to-noise ratio of the energy distribution curve are 29.14 (which does not meet the condition max(temp) / mean(temp)<9), so it enters the high SNR mode.
[0069] 2. The final threshold of 79.189 is obtained by calculating ThreCoe=mean(temp) through the threshold calculation.
[0070] 3. By using temp(Pos)>threshold, all points exceeding the threshold were successfully obtained. The points exceeding the threshold and the azimuth integral projection results are compared as follows: Figure 4 As shown.
[0071] 4. Obtain the target's starting position 237 and ending position 285 by using the distance to the point that exceeds the threshold.
[0072] 5. Following steps 1-4, perform the integral projection on the range direction to obtain the points that exceed the threshold in the azimuth direction. The points that exceed the threshold in the azimuth direction and the integral projection results are as follows: Figure 5 As shown, the target's orientation is obtained, with the starting position 247 and the ending position 272 to be captured.
[0073] Finally, it should be noted that the accompanying drawings of the embodiments disclosed in this invention only involve the structures involved in the embodiments disclosed in this invention. Other structures can refer to the general design. In the absence of conflict, the same embodiment and different embodiments of this invention can be combined with each other.
[0074] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. A method for rapid target extraction in ISAR images, characterized in that, include: Receive ISAR images, sum the gray values of the ISAR images in two directions, and generate integral projection vectors in two directions; Calculate the maximum and average gray values in the integral projection vectors of two directions, calculate the signal-to-noise ratio using the maximum and average gray values, and calculate the threshold coefficient based on the signal-to-noise ratio. The grayscale threshold value is calculated based on the threshold coefficient and the average grayscale value. The gray value of each pixel in the integral projection vector is compared with the gray value threshold. Pixels that do not meet the gray value threshold are removed to obtain the pixels and distance coordinates of the target image. The gray values of the pixels in the target image are summed along the distance direction to obtain the horizontal and vertical coordinates of the target image.
2. The method for rapid target capture in ISAR images as described in claim 1, characterized in that: The signal-to-noise ratio is calculated as follows: Calculate the ratio of the maximum grayscale value to the average grayscale value, and determine if it is less than 9. If it is, set the signal-to-noise ratio to 1.5; otherwise, set the signal-to-noise ratio to 1.
3. The method for rapid target capture in ISAR images as described in claim 1, characterized in that: The grayscale threshold is calculated by multiplying the signal-to-noise ratio by the average grayscale value.
4. The method for rapid target capture in ISAR images as described in claim 1, characterized in that: Set distance values in two directions respectively, count the coordinates of discrete points that exceed the grayscale threshold, and then compare the coordinates of the discrete points with the distance values in both directions. If the distance values in both directions are satisfied, the discrete point is determined to belong to the target image; otherwise, the discrete point is deleted.
5. A rapid target acquisition system for ISAR images, employing the method described in any one of claims 1-4, characterized in that, include: The image preprocessing module is used to receive ISAR images, sum the gray values in the azimuth direction of the ISAR images, and generate an integral projection vector. The signal-to-noise ratio calculation module is used to calculate the maximum and average gray values in the integral projection vector, calculate the signal-to-noise ratio using the maximum and average gray values, and calculate the threshold coefficient based on the signal-to-noise ratio. The grayscale threshold calculation module is used to calculate the grayscale threshold value based on the threshold coefficient and the average grayscale value. The target image determination module is used to compare the gray value of each pixel in the integral projection vector with the gray value threshold, and remove pixels that do not meet the gray value threshold to obtain the pixels and distance coordinates of the target image. The coordinate determination module is used to sum the gray values of the pixels in the target image along the distance direction to obtain the horizontal and vertical coordinates of the target image.
6. The ISAR image target rapid capture system as described in claim 5, characterized in that: In the signal-to-noise ratio calculation module, the ratio of the maximum grayscale value to the average grayscale value is calculated, and it is determined whether it is less than 9. If it is, the signal-to-noise ratio is set to 1.5; otherwise, the signal-to-noise ratio is set to 1.
7. The ISAR image target rapid capture system as described in claim 5, characterized in that: In the grayscale threshold calculation module, the signal-to-noise ratio is multiplied by the average grayscale value to obtain the grayscale threshold value.
8. The ISAR image target rapid capture system as described in claim 5, characterized in that: In the grayscale threshold calculation module, distance values in two directions are set respectively. The coordinate values of discrete points that exceed the grayscale threshold are counted. Then, the coordinate values of the discrete points are compared with the distance values in the two directions respectively. If the distance values in the two directions are satisfied, the discrete point is determined to belong to the target image; otherwise, the discrete point is deleted.