摘要：In this paper, we focus on the extraction problem of the target motion trajectory and make deep research. For Euclidean distance of SIFT feature matching algorithm is not adaptively adjustable, a improved SIFT feature matching algorithm based on multi-objective optimization is proposed. The optimization model is established to reduce mismatch rate, which consider Euclidean distance between correlation coefficient and feature point as the goal function and the confidence degree is taken as the constraint condition. Finally, The experiment shows that the method proposed in this paper can accurately achieve the measurement of the target motion trajectory, and is of strong adaptability and reliable
关键词：Video Image;Target Track;SIFT Feature Matching Algorithm;Optimization Model
