CN-121982346-A - Infrared target tracking method and system based on shielding judgment and UKF track prediction

CN121982346ACN 121982346 ACN121982346 ACN 121982346ACN-121982346-A

Abstract

The invention discloses an infrared target tracking method and system based on shielding judgment and UKF track prediction, which comprise the steps of acquiring an infrared target image under a complex background, selecting targets in the infrared target image by a frame, initializing a tracker, judging whether the targets of a current frame are in a shielding state or not based on a multi-feature shielding judgment mechanism to obtain a judgment result, modifying UKF prediction weights based on the judgment result to obtain predicted and corrected position information, and updating the positions and scales of the targets by the tracker according to the judgment result, the predicted and corrected position information, the scale step information of the current tracker and a scale polling mechanism.

Inventors

PENG HONGMING
LI JIACAN
WAN ZHONGNAN
WANG MINGYUAN
LI XINYUE

Assignees

西安建筑科技大学

Dates

Publication Date: 20260505
Application Date: 20260113

Claims (10)

1. An infrared target tracking method based on shielding judgment and UKF track prediction is characterized by comprising the following steps: acquiring an infrared target image under a complex background; Selecting a target in the infrared target image by a frame, and initializing a tracker; Judging whether the target of the current frame is in a shielding state or not based on a multi-feature shielding judging mechanism to obtain a judging result; based on the judgment result, modifying UKF prediction weight to obtain predicted and corrected position information; and according to the judgment result, the predicted and corrected position information, the scale step length information of the current tracker and the scale polling mechanism, carrying out position updating and scale updating of the target through the tracker.
2. The infrared target tracking method based on occlusion judgment and UKF trajectory prediction according to claim 1, wherein the acquiring the infrared target image in the complex background further comprises: And carrying out format conversion and standardization on the infrared target image under the complex background.
3. The infrared target tracking method based on occlusion judgment and UKF trajectory prediction according to claim 1, wherein the process of judging whether the target of the current frame is in an occlusion state based on the multi-feature occlusion judgment mechanism is as follows: calculating a final composite score The method comprises the following steps: Wherein, the Normalized peak weights; normalized PSR weights; normalized peak significance weights; is the normalized cross correlation significance; For normalized peak sidelobe ratio PSR, wherein Performing normalization processing, wherein x0=20, k=0.15; The peak value of the normalized response graph; When (when) When the target is judged to be in a non-shielding state, and when And judging that the target is in a shielding state.
4. The infrared target tracking method based on occlusion judgment and UKF trajectory prediction according to claim 3, wherein peak sidelobe ratio PSR is: Wherein, the The pixel intensity mean value of the sidelobe area; is the sidelobe region pixel intensity standard.
5. The infrared target tracking method based on shielding judgment and UKF track prediction according to claim 1, wherein when the target is in a non-shielding state, a target response confidence map is generated through image feature matching, a weighted average method is adopted to fuse the current target response confidence map with historical track information, smooth position updating is achieved, and a current frame scale polling result and a previous 3 frame scale change trend are combined, so that scale parameters of the target are dynamically adjusted through an exponential moving average algorithm.
6. The infrared target tracking method based on shielding judgment and UKF track prediction according to claim 1, wherein when the target is in a shielding state, a stagnation model is updated, a tracking state is recorded, and UKF weight proportion is increased when the position of the next frame is predicted.
7. An infrared target tracking system based on shielding judgment and UKF track prediction, which is characterized by comprising: the acquisition module is used for acquiring an infrared target image under a complex background; The frame selection module is used for selecting targets in the infrared target image in a frame mode and initializing a tracker; the judging module is used for judging whether the target of the current frame is in a shielding state or not based on the multi-feature shielding judging mechanism to obtain a judging result; the prediction module is used for obtaining position information of UKF prediction correction according to the judgment result; And the updating module is used for updating the position and the scale of the target through the tracker according to the judging result, the predicted and corrected position information, the scale step information of the current tracker and the scale polling mechanism.
8. The infrared target tracking system based on occlusion determination and UKF trajectory prediction as claimed in claim 7, wherein the process of determining whether the target of the current frame is in an occlusion state based on the multi-feature occlusion determination mechanism is as follows: calculating a final composite score The method comprises the following steps: Wherein, the Normalized peak weights; normalized PSR weights; normalized peak significance weights; is the normalized cross correlation significance; For normalized peak sidelobe ratio PSR, wherein Performing normalization processing, wherein x0=20, k=0.15; The peak value of the normalized response graph; When (when) When the target is judged to be in a non-shielding state, and when And judging that the target is in a shielding state.
9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the processor, when executing the computer program, implements the steps of the infrared target tracking method based on occlusion decisions and UKF trajectory prediction as claimed in any of claims 1-6.
10. A computer readable storage medium storing a computer program, wherein the computer program when executed by a processor implements the steps of the infrared target tracking method based on occlusion determination and UKF trajectory prediction as claimed in any one of claims 1-6.

Description

Infrared target tracking method and system based on shielding judgment and UKF track prediction Technical Field The invention belongs to the technical field of tracking, and relates to an infrared target tracking method and system based on shielding judgment and UKF track prediction. Background Infrared imaging techniques rely on imaging infrared radiation of a target, directly reflecting the difference in intensity of the target and background radiation. Under all-weather conditions, the unique imaging technology has the advantage of no visible light imaging. Infrared target tracking is the tracking of unmanned aerial vehicles, or other targets using infrared imaging techniques. The method is widely applied to civil fields such as robots, safety monitoring, urban management and the like, and military fields such as unmanned aerial vehicle searching reconnaissance, missile guidance and the like. Therefore, extensive research into infrared target tracking technology is necessary. Infrared target tracking aims to continuously locate and lack texture features, dynamic thermal features, low signal-to-noise ratio, targets that exhibit complex motion patterns from satellite video sequences or from the perspective of the drone (e.g., aircraft, drone, etc.). It mainly solves problems such as lack of obvious object features due to resolution limitations, strong background disturbances (e.g. clouds, surface textures) and non-linear movements (fast movements of objects). Due to the specificity of infrared target tracking, as well as the complexity and diversity of the environment, infrared target tracking faces challenges such as light source variation, background clutter, dimensional variation, occlusion, and the like. With the progress of technology, the performance of infrared target tracking is continuously improved. The performance of the tracking algorithm is obviously improved under the scene of light source and scale change, however, the problem of target tracking failure still exists in the occlusion scene. Therefore, it is important to improve the anti-occlusion performance of the infrared target tracking algorithm. When an infrared target is shielded, the tracking algorithm has the following adverse effects that ① is easily interfered by conditions such as deformation, background blurring and the like of the target in the target tracking process, the error judgment is that the target is shielded, the tracking result is abnormal, ② infrared target information is lost due to shielding of a cloud layer, a house and other obstacles, algorithm missing detection and target losing are caused, ③ infrared target information completely disappears under the condition of complete shielding, so that the target is difficult to capture again by the algorithm in a drifting state after the target reappears, and the accuracy of infrared target tracking is higher to be improved based on the problems. Disclosure of Invention The invention aims to overcome the defects of the prior art, and provides an infrared target tracking method and an infrared target tracking system based on shielding judgment and UKF track prediction. In order to achieve the above purpose, the invention discloses an infrared target tracking method based on shielding judgment and UKF track prediction, which comprises the following steps: acquiring an infrared target image under a complex background; Selecting a target in the infrared target image by a frame, and initializing a tracker; Judging whether the target of the current frame is in a shielding state or not based on a multi-feature shielding judging mechanism to obtain a judging result; based on the judgment result, modifying UKF prediction weight to obtain predicted and corrected position information; and according to the judgment result, the predicted and corrected position information, the scale step length information of the current tracker and the scale polling mechanism, carrying out position updating and scale updating of the target through the tracker. Further, the acquiring the infrared target image under the complex background further includes: And carrying out format conversion and standardization on the infrared target image under the complex background. Further, the process of judging whether the target of the current frame is in an occlusion state based on the multi-feature occlusion judgment mechanism is as follows: calculating a final composite score The method comprises the following steps: Wherein, the Normalized peak weights; normalized PSR weights; normalized peak significance weights; is the normalized cross correlation significance; is the normalized peak sidelobe ratio PSR, wherein. Using Performing normalization processing, wherein x0=20, k=0.15; The peak value of the normalized response graph; When (when) When the target is judged to be in a non-shielding state, and whenAnd judging that the target is in a shielding state. Further, the peak sidelobe ratio PSR is: Wherein,