Complex Background For The Rapid Detection And Tracking Of Targets

In recent years, computer vision recognition and digital image processing developed very rapidly, and image recognition and target tracking are very active in this field. Target tracking technology is very widely used in robot navigation, medical image analysis, intelligent monitor systems and video image compression and transmission. However, due to greatly changing environment, the present moving target detection algorithms are more complicated, time consuming and the lack of robustness, and suffered a lot of restrictions in the real deal especially in real-time applications. This paper focuses on rapid target detection and tracking in the complex background, including moving objects modeling and detection on the background and related problems, respectively the moving target detection simulation with the method of background subtraction, and CamShift algorithm on a single target tracking human motion simulation, and the using of Kalman filtering techniques in tracking movement of the human body which is blocked.This paper discusses the background difference method on the identification and tracking moving targets, and builds the model of the background, and then use background subtraction to moving object and background, through the gray scale processing, with the white prospects and black background.Secondly, CamShift algorithm is programmed based on the MeanShift algorithm and a single moving target tracking is simulated, besides the track picture is provided.Finally, this paper studies the separation of moving object occlusion problem, with Kalman filter proposed for the target state prediction and estimation, thereby increasing the target block and determine the reliability of the separation. With Kalman filtering and CamShift combination, this paper presents a full track design for moving target with block and provids simulation experiment in accordance of this algorithm, and the results are satisfactory in tracking the moving targets.