1 简介
Object tracking is a well studied fifield in computer vision. The goal is to use a sensor suchas a camera or depth sensor to track an object in environments which may include occlusionor low lighting. The state of the art algorithms in this fifield uses particle fifilters to maintainestimates of an objects location and speed. These algorithms take their cue from researchstarted in the 1980s.One early work which had a large impact was Ted Broida’s “Estimation of Object MotionParameters from Noisy Images.” His work was one of the fifirst to address the problem ofobject tracking using a stochastic method such as the Kalman Filter. His method utilized theIterative Extended Kalman Filter (IEKF) as a recursive estimation procedure to determinean object’s motion parameters over