Intelligent Target Detection and Tracking Algorithm for Martial Arts Applications

Moving object detection and tracking is the basis key technology for intelligent video surveillance, human-computer interaction, mobile robot vehicle visual navigation, industrial system, other applications. It has important applications in monitoring, so on. Intelligent can greatly facilitate monitoring of countless targets. This paper adopts random motion model to describe state target. Based on wireless communication information security, this studies security system martial arts target algorithm, basic idea mean-shift algorithm use gradient climbing probability density find local optimal. uses background subtraction method based vibe obtain binary image foreground object; shadow used remove image; haar-like (Haar) feature selected as detection, value rectangular area quickly calculated adjacent area. The difference between features, final result image, then analysis. Experimental data shows that time consumed by less than 20 ms, which meet real-time requirements ordinary systems. average processing hybrid modeling 62.8 rate 15.92 frames/SEC. results show improves utilization particles, reduces complexity, degradation particle filter.