Research And Realization On Detection And Tracking Of Objects Based On Video-stream

Object detection and tracking based on video-stream, which includes up-to-date technologies like image processing, pattern recognition, artificial intelligence, automatic control, computer science, etc, is becoming an important domain in the area of computer image processing. But currently, the detection and tracking technologies are not perfect in detecting and tracking the moving object in complex environment or in large area with multi-targets. The methods still need to be improved. According to current status, follow research has been done in the paper.(1) For the moving object detection, an adaptive background extraction method is proposed and the real-time background updating is also completed to decrease the dependency on the changes caused by the light or the environment. For the "holes" generated by the moving object in the subtraction image, the paper use morphological filter to deal with the results of the object segmentation, which eliminates background noise and the impact of disturbance.(2) For the moving object tracking, it is based on moving object detection. The main aim of tracking is to get the moving features of moving object, including movement trajectory and so on. The paper proposes a regional matching tracking method based on Kalman Filtering. Though Kalman filtering method, we could predict the next position of the object in the object zone and estimate the object possible location of the next moment, then we could search for matching object and create correspondence between the object frame in this small area.In order to validate the algorithm, we accomplished it in a filter with the DirectShow SDK developing environment, and made certain moving objects tracing tests. Results show that the algorithm in this paper could get certain satisfactory results in different environment.