Human Body Motion Tracking and Prediction Based on Buchberger Algorithm

Abstract

To ensure the real-time and accuracy of the human tracking system, it is necessary to predict the position of the target at the next moment according to the current trajectory of the target. This paper presents an accurate detection, extraction and tracking algorithm for human motion targets in video sequences. The algorithm uses the inter-frame difference threshold (TIFD) method to quickly and accurately detect and extract the target. The Buchberger algorithm is used to predict the area which may be in the next moment of the target, and the search scope of the target is reduced. Making full use of the results of running target detection can improve the target’s matching efficiency and tracking speed. At the same time, the corresponding experimental results are given. The results show that the method is more practical and can meet the basic requirements of human motion analysis.