A Novel Motion Object Capture And Tracking Algorithm Based On The Soldiers Shooting System

Motion object capture and tracking of the video stream is one of the hotspots among the research subjects. It can realize the figure capture and trajectory tracking of the motion objects with the combination of many advanced subject technologies such as digital image processing, pattern recognition, artificial intelligence and computer control.Generally speaking, motion object capture algorithm is based on the static background. However, the background is comparatively complicated in the practical application under the influence of illumination variation and camera shake. The soldiers shooting system discussed in this paper is to capture and track soldiers’moving figures in the complex background.In this soldiers shooting system, the shooting venue is divided into two parts:the right part and the left part. Two industrial view cameras are recording the soldiers’motion and shooting in front of the shooting venue. Meanwhile, the computer can trace and track the moving targets at the same time making use of the image sequences from the video. Finally, the systems will determine the winner of the shooting contest on the basis of motion capture and tracking result.The key point of the motion capture and tracking algorithm is the contradiction of the algorithm accuracy and real-time performance. To be more specific, the contradiction of the large-sized original image and relatively slow processing speed is the main issue of this system. On the basis of the traditional background subtraction approach, an optimized algorithm which narrows the scope of capture zone is proposed. According to the characteristic of the moving objects appear in the specific location, a’capture zone’is used to capture and track the targets. The system only search the pixels in the capture zone rather than the whole image which will definitely reduce the calculation amount. By means of Canny edge detection, morphological processing and de-noising technology, the’capture zone’ will be narrowed down to further accelerate the algorithm.This algorithm proves to be favorable shorten time of searching all zero gray value pixels on the condition that it ensures the accuracy of capture motion object. Furthermore, the region growing method is used to merge the object of adjacent during the tracking process. At last, Kalman filter is used as a tracking filter in the tracking process of the system which can find the best match with the object model candidate object location by Kalman filter in the search area and match it with the object template of the current frame. Experiments on video sequences of the soldiers shooting, the algorithm has the advantage of high-speed, anti-noise and it can detect the motion object.