Dim Moving Point Target Detection & Tracking Based On Multi-source IR Image Sequences

In this paper, the study on DBT detection & tracking algorithm for a dim nonlinear moving point target in infrared image sequence is based on the former researches, and then considering the characteristics of infrared similar-type imaging devices from different perspectives, it will case differential background and movement locus in the multi-source infrared image sequences, the attempts which are multi-source image sequences fusion tracking and multi-vision tracking & orientation technology about a dim moving point target have been processed.Adopt MATLAB simulating test: a dim nonlinear moving point target tracking algorithm in infrared image sequence which bases on Gaussian particle filter largely improves tracking precision and has a good real-time performances; compared to the single infrared image sequence, experimental analysis and results show that multi-source measurements information fusion tracking algorithm has high accuracy and good stability, simultaneity, avoids losing the single sequence track; the study on a dim moving point target's particle filter tracking and orientation bases on binocular, experimental results show that after estimating the density function's parameters, the algorithm which distributes the weight to particle according to PDF advances the precision of image tracking, the method which adapts linear triangulation principle to extract the target's spatial localization is effective.