A method for fast screening and eliminating false targets in multi-channel SAR images

Abstract

The invention provides a method for fast screening and eliminating false targets in a multi-channel SAR image, belonging to the technical field of radar signal processing. The specific process of the method is as follows: first, calculating the distance between false targets on both sides of real targets, sorting all targets detected in the SAR image according to azimuth, calculating the distance interval between two adjacent targets, and grouping the targets based on the distance, the distance interval and a fault tolerance threshold; secondly, making a uniform judgment on the attributes of the same group of targets, and removing targets that do not have the same attributes; finally, for each group of targets, eliminating false targets based on the scattering characteristics of the targets, so as to realize fast screening and eliminating false targets in the multi-channel SAR image.

Description

Technical Field

[0001] The invention belongs to the technical field of radar signal processing, and in particular relates to a method for quickly screening and eliminating false targets in multi-channel SAR images. Background Art

[0002] Synthetic aperture radar (SAR) is an active microwave remote sensing device with all-weather and all-day earth observation capabilities. As the application scope of SAR in military and civilian fields becomes wider and wider, the requirements for SAR swath width and resolution in various fields are getting higher and higher. Azimuth multi-channel technology has well solved the contradiction between traditional SAR resolution and swath width, realizing high-resolution wide swath imaging, and has become an important technical means for the next generation of spaceborne SAR to achieve high-resolution wide swath imaging. However, in actual systems, factors such as the amplitude and phase inconsistency of the radar system's multi-channel receiving channels, the instability of satellite attitudes, and the rapid movement of ships will cause amplitude and phase errors between channels, leading to the introduction of false targets.

[0003] The existence of SAR multi-channel false targets poses a huge challenge to SAR ship detection. These false targets are very bright relative to the sea surface, and will be misdetected when they appear in low-scattering areas such as the sea surface, resulting in false alarms. Although there is a certain attenuation in brightness between multi-channel false targets and corresponding real targets, there is no clear relationship between brightness and darkness between them and other non-corresponding real targets, so they cannot be identified based on absolute brightness. What is more serious is that during the SAR imaging process, the movement of the ship causes the change of the Doppler parameters, which in turn causes the mismatch of the azimuth matched filter, so the ship target will produce different degrees of defocus in the SAR image. How to distinguish the false targets caused by multi-channels from the real defocused targets in the full-aperture image is of great significance to reducing the false alarm rate in SAR ship detection. However, only a few scholars have conducted relevant research in this regard. Ma Lin et al. proposed a SAR multi-channel false target identification method based on sub-aperture and full-aperture feature learning to solve this problem. While using a deep convolutional neural network to extract full-aperture features, the complex SAR image is decomposed into sub-apertures and the sub-aperture features are extracted using a stacked convolutional autoencoder network. The identification information contained in the sub-aperture features is used to distinguish between real targets and multi-channel false targets. However, this method uses two deep convolutional networks for feature extraction, which is very computationally intensive and difficult to implement, especially when resources are limited, such as onboard spacecraft. Summary of the invention

[0004] In view of this, the present invention provides a method for fast screening and eliminating false targets in multi-channel SAR images, which can achieve accurate elimination of false targets.

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

[0006] A method for fast screening and eliminating false targets in multi-channel SAR images. The specific process is as follows:

[0007] Calculate the distance between the false target and the real target;

[0008] Sort all targets detected in the SAR image by azimuth and calculate the distance interval between two adjacent targets;

[0009] Grouping the targets based on the spacing, the distance interval and the error tolerance threshold;

[0010] Make a uniform judgment on the attributes of the same group of targets and remove targets that do not have the same attributes;

[0011] For each group of targets, false targets are eliminated based on the scattering characteristics of the targets.

[0012] Furthermore, when the present invention groups the targets, the group containing only one target is removed.

[0013] Furthermore, the present invention groups targets based on the spacing, the distance interval and the tolerance threshold as follows: if |d-n1|≤l is satisfied, two adjacent targets are grouped into one group, wherein l1 is the tolerance threshold, and n1n1 is the spacing between the false target on both sides of the real target.

[0014] Furthermore, the interval n1 at which the false targets of the present invention are equally spaced on both sides of the real target is:

[0015]

[0016] Among them, N a单补零 is the number of points after zero padding in single-channel azimuth, m is the number of gates at the target distance, R min is the minimum slope distance, f s is the sampling frequency, T p is the pulse width, R ref is the reference slant distance and c is the speed of light.

[0017] Furthermore, the uniformity judgment of the attributes of the same group of targets described in the present invention is: uniformity judgment is performed based on two attributes: target direction and target size.

[0018] Furthermore, the target direction and target size of the present invention are determined according to the following method:

[0019] Cut out a slice of set size with the target center point as the center;

[0020] Perform Otsu threshold segmentation on the slice containing the target to obtain a segmented binary image;

[0021] The minimum bounding rectangle of the target area of ​​the binary image is calculated to obtain the target direction and target size.

[0022] Furthermore, after removing targets that do not have the same attribute, if there is only one target in a group, the group is removed.

[0023] Furthermore, the specific process of eliminating false targets based on the scattering characteristics of the target described in the present invention is: for the targets in each group, mean and variance statistics are performed respectively, wherein the mean reflects the brightness information of the target, and the variance reflects the degree of scattering of the target, the target with the largest mean and variance is considered to be the real target, and the other targets are judged as false targets, thereby achieving the elimination of false targets.

[0024] Beneficial Effects

[0025] First, the present invention first groups the detection results into suspected false target groups according to the characteristic that false targets are symmetrically distributed at equal intervals on both sides of the real target, and then further eliminates the targets based on the target attributes. This method can realize the rapid screening and elimination of false targets in SAR images, thereby achieving the purpose of reducing the false alarm rate.

[0026] Secondly, the present invention makes identity judgment based on attribute features such as target orientation and length, and screens and removes suspected false target groups, which can quickly remove targets that do not belong to the same attribute, thereby avoiding grouping errors caused by over-dense grouping.

[0027] Thirdly, the present invention eliminates false targets by using the difference in scattering characteristics between false targets and real targets, thereby achieving rapid screening and elimination of false targets in SAR images. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 Schematic diagram of false targets and real targets.

[0029] Figure 2 Schematic diagram of the distribution of false targets and real targets.

[0030] Figure 3 Schematic diagram of false target elimination results. DETAILED DESCRIPTION

[0031] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0032] It should be noted that the following embodiments and features in the embodiments may be combined with each other in the absence of conflict; and, based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in the field without making any creative work are within the scope of protection of the present disclosure.

[0033] It should be noted that various aspects of the embodiments within the scope of the appended claims are described below. It should be apparent that the aspects described herein may be embodied in a wide variety of forms, and any specific structure and / or function described herein is merely illustrative. Based on the present disclosure, it should be understood by those skilled in the art that an aspect described herein may be implemented independently of any other aspect, and two or more of these aspects may be combined in various ways. For example, any number of aspects described herein may be used to implement the device and / or practice the method. In addition, other structures and / or functionalities other than one or more of the aspects described herein may be used to implement this device and / or practice this method.

[0034] The present invention proposes a method for fast screening and eliminating false targets in multi-channel SAR images based on position information, and the steps are as follows:

[0035] Step 1: Obtain multi-channel SAR images, imaging related parameters and detection results, and the target to be detected, such as Figure 1 As shown, they are located at the upper left corner and lower right corner of the four-boundary coordinates in the multi-channel SAR image.

[0036] Step 2: Calculate the distance between the false target and the real target; sort all the targets detected in the SAR image according to the azimuth, and calculate the distance interval between two adjacent targets; group the targets based on the distance, the distance interval and the fault tolerance threshold.

[0037] The basis for grouping in this step is: According to the imaging characteristics of multi-channel false targets, false targets will be distributed at equal intervals on both sides of the real target, such as Figure 2 As shown, if the target-to-target interval is within the tolerance threshold, they are grouped together.

[0038] When this step is implemented, the interval n1 between the targets can be expressed as:

[0039]

[0040] Among them, N a单补零 is the number of points after zero padding in single-channel azimuth, m is the number of gates at the target distance, R min is the minimum slope distance, f s is the sampling frequency, T p is the pulse width, R ref is the reference slant distance and c is the speed of light.

[0041] Detect all targets and group the detection results that meet the interval conditions. The specific steps are as follows:

[0042] (1) Arrange the target center point coordinates of the same scene image detection results in ascending order of azimuth.

[0043] (2) Select the first target as the first group, and then calculate the distance interval between the first target and the adjacent second target. Assuming that the coordinates of the center points of the first and second targets are (x1, y1) and (x2, y2) respectively, the distance interval is:

[0044]

[0045] If satisfied:

[0046] |d-n1|≤l

[0047] Among them, l is the fault tolerance threshold, the second target is merged into the first group, if the detection result does not meet the fault tolerance threshold, a new group is started.

[0048] (3) Similarly, the distance interval between each subsequent target and the last target in each group is calculated and conditional judgment is performed. If the fault tolerance threshold is met, it is merged into the corresponding group. If not, a new group is created until all targets are grouped.

[0049] (4) Remove the target group with only one target, because when there is only one target in the group, it means that the target is the real target.

[0050] Step 3: Make a uniform judgment on the attributes of the same group of targets and remove targets that do not have the same attributes;

[0051] Because there are multiple real targets in the image, in order to avoid grouping errors caused by too dense targets, the direction, size and other attributes of the target will be calculated for each group of targets, and the identity will be judged. The target coordinates that do not meet the identity will be removed from the group. Therefore, this step further groups the false targets according to the identity.

[0052] The specific process of this step is as follows:

[0053] (1) According to the detection result information, a 200*200 or 400*400 slice is cut out with the target center coordinate as the center. The slice size depends on the image resolution.

[0054] (2) Perform Otsu threshold segmentation on the slice containing the target to obtain a segmented binary image.

[0055] (3) Calculate the minimum bounding rectangle of the target area in this binary image and obtain the target's direction, size and other attributes.

[0056] (4) Perform identity judgment on the attributes of the same group of targets in step 2, and remove targets that do not belong to the same attributes from the group.

[0057] (5) Remove the target group with only one target, because there is only one target in the group, which means that the target is the true target.

[0058] Step 4: For each group of targets, false targets are eliminated based on the scattering characteristics of the targets.

[0059] In this step, each group of targets obtained after step three is considered to contain one real target and multiple false targets in the same group, and the false targets are identified based on the differences in brightness and scattering characteristics between the false targets and the real targets of the same target.

[0060] The specific process of removing false targets based on the target's scattering characteristics is as follows: for each target in each group, the mean and variance statistics are performed (based on the target's four-boundary coordinate range), where the mean reflects the target's brightness information and the variance reflects the target's scattering degree. The target with the largest mean and variance is considered to be the real target, and the other targets are judged as false targets, thereby achieving the removal of false targets, such as Figure 3 shown.

[0061] 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 rapid screening and elimination of false targets in multi-channel SAR images. The specific process is as follows: Calculate the distance between the false target and the real target; Sort all targets detected in the SAR image by azimuth and calculate the distance interval between two adjacent targets; Grouping the targets based on the spacing, the distance interval and the error tolerance threshold; Make a uniform judgment on the attributes of the same group of targets and remove targets that do not have the same attributes; For each group of targets, false targets are eliminated based on the target's scattering degree and brightness information; The distance n1 between the false target and the real target is: Among them, N a单补零 is the number of points after zero filling in single channel azimuth, m is the number of gates at the target distance, R min is the minimum slope distance, f s is the sampling frequency, T p is the pulse width, R ref is the reference slant distance, c is the speed of light; The targets are grouped based on the spacing, the distance interval and the tolerance threshold: if |d-n1|≤l is satisfied, two adjacent targets are grouped together, where l is the tolerance threshold, n1 is the spacing between the false target on both sides of the real target, and d is the distance interval.

2. According to the method of claim 1, the method is characterized in that: When grouping targets, remove groups that have only one target.

3. According to the method of claim 1, the method is characterized in that: The uniformity judgment of the attributes of the same group of targets is: uniformity judgment is performed based on two attributes: target direction and target size.

4. According to claim 3, a method for rapid screening and elimination of false targets in multi-channel SAR images is characterized in that: The target direction and target size are determined as follows: Cut out a slice of set size with the target center point as the center; Perform Otsu threshold segmentation on the slice containing the target to obtain a segmented binary image; The minimum bounding rectangle of the target area of ​​the binary image is calculated to obtain the target direction and target size.

5. According to claim 1, a method for rapid screening and elimination of false targets in multi-channel SAR images is characterized in that: After removing targets that do not have the same attributes, if there is only one target in the group, the group is removed.

6. According to the method of claim 1, the method is characterized in that: The specific process of eliminating false targets based on the scattering degree and brightness information of the target is as follows: for the targets in each group, mean and variance statistics are performed respectively, wherein the mean reflects the brightness information of the target and the variance reflects the scattering degree of the target, the target with the largest mean and variance is considered to be the real target, and the other targets are judged to be false targets, thereby achieving the elimination of false targets.

Images