Study And Implementation Of Multi-Target Tracking Algorithm

As a widely used technology, multi-target tracking from image sequences has been given attention by many experts. Although a number of algorithms have been proposed, there are still a lot of problems during practical application. Under these conditions, the aim of this thesis is to study image feature extraction and prediction tracking algorithms systematically, to explore and summarize an applicable multi-target tracking scheme, based on the available researches. Then, with a better understanding, we hope to make some improvements.This paper expatiates on the meaning, the theory foundation, actual technology and research status of multi-target tracking. A survey in multi-target tracking research field is applied in this paper at first. Then, on the basis of collecting and analyzing large data in multi-target tracking field, this paper concludes a model that is suitable for the engineering application. To resolve the question of tracking maneuvering targets over simple background, the theory of kalman filter and a cost function algorithm of multi-target tracking have been analyzed in depth. In this thesis, with the help of MATLAB simulating tool, it accomplished the tracking of two maneuvering targets and completed the algorithm emulation of every phase in multi-target tracking. From the satisfied result of simulation, the validity and the accuracy of multi-target tracking algorithm has been proved.In addition, a novel design and implementation of relay protection and testing system is also introduced in this paper. The system scheme and the software arithmetic are discussed in details. This system can measure and verify many technical parameters of various delay devices.