Target Tracking Based on Multi Feature Selection Fusion Compensation in Monitoring Video

摘要

This thesis is mainly targeted at self-adaptation adjustment in the search region: at first, design a staging predation space self-adaptation scale strategy bat algorithm (AP-RBA), and then, use AP-RBA algorithm to establish a target tracking strategy of optimized particle filter which can effectively solve two kinds of problems: (1) particle impoverishment phenomena produced in particle filter; (2) effective tracking targets based on few particles, thus simplifying complexity of particle filter, and then, adopt the criterion weight strategy to achieve maximum a posteriori and change of criterion weight to realize effective improvement of particle distribution and promote efficiency of particle filter process.