Study on the video image tracking technology based on the particle filter algorithm

The complexity of video images and the existence of movement mutability substantially reduce correlation between movements, and at the meantime result to large deviation of tracking results. Thus it is hard to track the video images accurately with traditional tracking algorithms. In order to solve this problem, the author proposes a new visual tracking technology based on particle filter image frames. The new algorithm introduces the randomly distributed movement mutability impact operator as the penalty factor, and at the meantime samplings video images with particle filter visual objects, so as to obtain the real-time tracking. The experimental results prove that the algorithm proposed in this paper avoids the delay of the traditional tracking algorithms, and realizes the real-time accurate tracking of targets.