Application of design of image tracking by combining SURF and TLD and SVM-based posture recognition system in robbery pre-alert system

This paper describes the design of an image tracking system that combines Speeded Up Robust Features (SURF) and Tracking-Learning-Detection (TLD), with a posture recognition system that is based on the Support Vector Machine (SVM), and includes image tracking, foreground detection and posture recognition. Image tracking, which combines the SURF and the TLD algorithms, starts by detecting the postures of a specific person with SURF, before tracking the object using TLD. Foreground detection uses the Codebook background subtraction algorithm to acquire foreground images and mends them by post-processing. Lastly, posture recognition applies SVM to determine human body postures. This article embodies such a system by applying it in a robbery pre-alert system.