CN-122023844-A - Sea surface target association method based on feature topological graph matching

CN122023844ACN 122023844 ACN122023844 ACN 122023844ACN-122023844-A

Abstract

The application belongs to the technical field of sea surface target association, and particularly discloses a sea surface target association method based on characteristic topological graph matching. The method comprises the steps of firstly generating a characteristic topological graph of targets to be associated based on target sea surface distribution detected by a detector, forming a target pair to be associated by two targets to be associated detected by different detectors, enabling the number of reference target pairs meeting neighborhood coincidence to be the largest in the two characteristic topological graphs of the target pair to be associated through rotation and translation, then evaluating the matching degree of the reference target pair by the distance between the two reference targets in the reference target pair, evaluating the matching degree of the target pair to be associated by the matching degree of all the reference target pairs, and finally iteratively evaluating the matching degree of all the target pairs to be associated and associating the target pairs to be associated with the optimal matching degree into the same target. The sea surface target correlation method does not depend on absolute position information of the target, and has stronger adaptability and robustness.

Inventors

Wu hanbao
LIAO MINGLI
ZHOU YUTING
WANG HANG
YANG YONGGANG

Assignees

中国船舶集团有限公司第七〇九研究所

Dates

Publication Date: 20260512
Application Date: 20260122

Claims (10)

1. The sea surface target association method based on feature topological graph matching is characterized by comprising the following steps of: Generating a characteristic topological graph of a target to be associated based on the target sea surface distribution detected by the detector, wherein the target to be associated is used as the center of the characteristic topological graph, and surrounding reference targets are mapped on the characteristic topological graph according to the target sea surface distribution; Two targets to be associated detected by different detectors form a target pair to be associated, and the number of reference target pairs meeting neighborhood coincidence is the greatest in the two characteristic topological graphs of the target pair to be associated through rotation and translation; Estimating the matching degree of the reference target pair by the distance between two reference targets in the reference target pair, and estimating the matching degree of the target pair to be associated by the matching degree of all the reference target pairs; And simultaneously, associating the reference target pairs with the same target in the two feature topological graphs of the target pairs to be associated, which realize neighborhood coincidence.
2. The sea surface target association method according to claim 1, wherein the number of reference target pairs meeting neighborhood coincidence is the largest in the two feature topological graphs of the target pairs to be associated through rotation and translation, specifically: (1) Aligning centers of the first feature topology map and the second feature topology map; (2) If a first reference target in the first characteristic topological graph and a second reference target in the second characteristic topological graph meet neighborhood coincidence, the first reference target and the second reference target are regarded as a datum reference target pair; (3) The second characteristic topological graph is rotated and translated to enable the datum reference target pairs to completely coincide, and at the moment, if a third reference target in the first characteristic topological graph and a fourth reference target in the second characteristic topological graph meet neighborhood coincidence, the third reference target and the fourth reference target point are regarded as preselected reference target pairs; (4) Reselecting the base reference object pairs and returning to the step (3) until all the base reference object pairs are traversed; (5) The rotation and translation that yields the greatest number of pre-selected reference target pairs is considered as the optimal registration, with all pre-selected reference target pairs obtained in the optimal registration being the reference target pairs.
3. Sea surface target correlation method according to claim 1 or 2, characterized in that the meeting neighborhood overlap is specifically: The characteristic topological graph is divided into a plurality of areas at equal intervals by a preset polar angle step length and a polar diameter step length, and each area is provided with a unique code according to the actual position; if the codes of the region where the first reference target is located in the first characteristic topological graph and the codes of the region where the second reference target is located in the second characteristic topological graph are neighborhood codes, judging that the first reference target and the second reference target meet neighborhood coincidence; In the feature topology, the codes of two regions having at least one common point are the neighborhood codes, and the codes of the same region are also the neighborhood codes.
4. A method of relating to sea surface targets according to claim 3, wherein the region is encoded as 。
5. The sea surface target correlation method according to claim 1, wherein the matching degree of the reference target pair is evaluated by the distance between two reference targets in the reference target pair, specifically, the mahalanobis distance between the two reference targets is calculated, and the smaller the mahalanobis distance is, the higher the matching degree of the reference target pair is.
6. The method according to claim 5, wherein the mahalanobis distance between the two reference targets is : ; Wherein, the Representing a characteristic topology Reference object in (2) , Representing a characteristic topology Reference object in (2) ; Representing a reference target And Is a difference vector of (a): ; Is the reference target And The difference in the direction of the X-axis, Is the reference target And Difference in Y-axis direction; The transpose of the matrix is represented, As an error covariance matrix: ; Is the measurement error of the detector in the X-axis direction, Is the measurement error of the detector in the Y-axis direction.
7. The sea surface target correlation method according to claim 1, wherein the matching degree of the target pairs to be correlated is evaluated by the matching degree of all reference target pairs, specifically, the matching degree of the target pairs to be correlated is evaluated by the matching degree average value of all reference target pairs.
8. The sea surface target association method according to claim 1 is characterized in that target pairs to be associated with optimal matching degree are associated to be the same targets, specifically, whether the matching degree of the target pairs to be associated with optimal matching degree exceeds a preset matching threshold is judged, and if so, the target pairs to be associated are associated to be the same targets.
9. The sea surface target correlation method of claim 8, wherein the preset matching threshold satisfies: ; The above expression, in the original assumption In the case where the target pairs to be associated are the same, i.e. the degree of matching of the target pairs to be associated Is greater than a preset matching threshold Probability of (2) Is equal to a significant level By setting a significant level Looking up a chi-square distribution table with the degree of freedom of 2 to obtain the preset matching threshold value 。
10. An electronic device comprising a memory, one or more processors; The memory is coupled to the one or more processors, the memory for storing computer program code, the computer program code comprising computer instructions; The one or more processors invoking the computer instructions to cause the electronic device to perform the method of any of claims 1-9.

Description

Sea surface target association method based on feature topological graph matching Technical Field The application belongs to the technical field of sea surface target association, and particularly relates to a sea surface target association method based on feature topological graph matching. Background The sea surface target correlation technology is a core link of modern sea surface information fusion, and has the core task of correctly judging whether target signals detected by a plurality of detectors from different platforms and at different times belong to the same real target or not, so that discrete observation points are fused into a continuous, stable and unified target track diagram, and a reliable real-time situation sensing basis is provided for command decision. The method is similar to a 'jigsaw master' on the sea surface, and the fragmented detection information is spliced into a complete sea surface situation panorama through the methods of position comparison, motion prediction, feature matching and the like. At present, the traditional sea surface target correlation method relies on absolute position information of a target, is extremely sensitive to systematic rotation and translation errors of a radar, and can be rapidly deteriorated when facing a complex environment with high false alarm and low detection rate. Meanwhile, the method has high calculation complexity, is difficult to effectively process high-dynamic scenes such as dense targets, cross tracks and the like, causes insufficient environment adaptability, and cannot meet the severe requirements of modern sea surface scheduling on high reliability and high instantaneity. Disclosure of Invention Aiming at the defects of the prior art, the application aims to provide a sea surface target association method based on characteristic topological graph matching, which aims to solve the technical problem that the existing sea surface target association technology is insufficient in environment adaptability. The first aspect of the application relates to a sea surface target association method based on characteristic topological graph matching, which comprises the following steps: Generating a characteristic topological graph of a target to be associated based on the target sea surface distribution detected by the detector, wherein the target to be associated is used as the center of the characteristic topological graph, and surrounding reference targets are mapped on the characteristic topological graph according to the target sea surface distribution; Two targets to be associated detected by different detectors form a target pair to be associated, and the number of reference target pairs meeting neighborhood coincidence is the greatest in the two characteristic topological graphs of the target pair to be associated through rotation and translation; Estimating the matching degree of the reference target pair by the distance between two reference targets in the reference target pair, and estimating the matching degree of the target pair to be associated by the matching degree of all the reference target pairs; And simultaneously, associating the reference target pairs with the same target in the two feature topological graphs of the target pairs to be associated, which realize neighborhood coincidence. Preferably, the number of reference target pairs meeting neighborhood coincidence is the largest in the two feature topological graphs of the target pairs to be correlated through rotation and translation, specifically: (1) Aligning centers of the first feature topology map and the second feature topology map; (2) If a first reference target in the first characteristic topological graph and a second reference target in the second characteristic topological graph meet neighborhood coincidence, the first reference target and the second reference target are regarded as a datum reference target pair; (3) The second characteristic topological graph is rotated and translated to enable the datum reference target pairs to completely coincide, and at the moment, if a third reference target in the first characteristic topological graph and a fourth reference target in the second characteristic topological graph meet neighborhood coincidence, the third reference target and the fourth reference target point are regarded as preselected reference target pairs; (4) Reselecting the base reference object pairs and returning to the step (3) until all the base reference object pairs are traversed; (5) The rotation and translation that yields the greatest number of pre-selected reference target pairs is considered as the optimal registration, with all pre-selected reference target pairs obtained in the optimal registration being the reference target pairs. Preferably, the neighborhood registration is satisfied specifically as follows: The characteristic topological graph is divided into a plurality of areas at equal intervals by a preset polar angle step leng