Abstract
An infrared small target detection and tracking method suitable for different scenes is proposed. For a given infrared image sequence, 1) firstly, the complexity of the scene is judged according to the number of Harris corners; 2) for the image sequence containing simple scenes, the method of local adaptive threshold segmentation and track association is used to complete the detection and tracking of multi-target; 3) for the image sequence containing complex scenes, SSD detector is used to detect the initial target, after the target is detected, SiamRPN++ tracker is started to track the target. The tracking result is judged in the tracking process. When the target is lost or wrong, the SSD detector is restarted immediately to capture the correct target position and fed it into the tracker to complete the reinitialization and correction. In the tracking process, different targeted strategies are adopted to deal with different situations, such as the target entering and leaving the field of view, the sudden change of the moving state, false alarm interference and so on, ensuring that the detection and tracking of the whole image sequence is completed efficiently and stably. The proposed method which provides solutions for different scenes can successfully realize the continuous and accurate positioning of multiple infrared targets in different scenes, improving the ability of the infrared detection system to solve practical problems. Especially for the complex scene, the neural network based intelligent detector and tracker complement each other to improve the detection and tracking efficiency of infrared dim small targets, and have strong robustness and generalization performance.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.