Moving Target Detection And Tracking Control Study Of Mobile Robot In Complex Environments

With the development of computer vision system and control technology, mobile robot attracted enough attention in the actual environment.For a mobile robot, realizing the control of the dynamic target detection and tracking is an important manifestation of intelligence, this technology has broad development in the field of automation technology, video surveillance and so on.The research of dynamic target detection and tracking control of mobile robot has become a hot topic detection and tracking, to achieve control of the dynamic target how to rely on the mobile robot is a challenging research. In this paper, the main research contents are as follows:First,in terms of moving target detection,Kim algorithm combined with inter-frame difference method,background subtraction and a Gaussian mixture background model, It includes that the background image is updating the mixed Gauss background model, the inter frame difference method and background subtraction method and the motion area were extracted from the morphological filtering and connectivity check and other steps, the final completion of motion target is extracted. The Kim algorithm can be real-time updating background, improve the segmentation accuracy of target area,he clear outline of the background image, under the case of other objects includes partial occlusion that illumination changes, the target color and background color similar the case of other objects,it complete to detect the target.Secondly,it analysis of the basic method, dynamic target tracking and their respective advantages and disadvantages, and then puts forward a kind of dynamic object tracking method based on improved Camshift algorithm.In the input video sequence images, dynamic target is moving quickly, it can set the smaller time interval between frames, change of state in two consecutive frames of dynamic target is small, the target displacement is small, it can be regarded as the target uniformly accelerated rectilinear motion in the direction of coordinate axis respectively, predicted by the acceleration the displacement equation of motion trajectory of the dynamic target, reduce the number of search window, adaptive search window size and location.Finally, in the experiment,through Visual Studio 2010 and Opencv2.4.10 simulation of the input video image sequences of moving targe,it test The effect of detecting and tracking;In the laboratory environment, the Kim algorithm and the improved camshift algorithm is transplanted to a mobile robot verify the reliability of the algorithm.