Research On Moving Target Detection And Tracking Technology In Classroom Video Recording And Broadcasting System

The emergence of intelligent video recording system is a revolution for the traditional video recording in the classroom teaching. It completely solves the problem of complexity and interference of artificial classroom shooting, so that the open class recorded closer to the true class effect. And it can combine with the Internet to achieve real-time broadcast of classroom teaching process, in order to broaden the scope of the classroom, and optimize the allocation of educational resources. In this context, this paper studies moving target detection and tracking algorithms in the classroom video recording and broadcasting system, mainly including the following three aspects:Firstly, this paper analyzes the actual needs of classroom video recording and broadcasting system and studies the existing target detection algorithm, then selects the moving target detection algorithm based on background subtraction method. On this basis, this paper focuses on the background updating, image thresholding, contour extraction and other related algorithms. And use the technology of contour defect repair to improve the contour extraction algorithm. Experimental results show that the proposed algorithm provides a good technical support to realize the recording and broadcasting system.Then, this paper analyzes the actual needs of classroom video recording and broadcasting system and chooses a moving target tracking algorithm based on compressed sensing. The basic idea is to use compressed sensing theory to extract the low-dimensional image feature, and then use the naive Bayes classifier for classification in order to achieve the target tracking. This paper uses the Kalman filter to further improve the existing tracking algorithm. Experiments show that the proposed algorithm with real-time tracking and robustness.Finally, from reality, we present a complete intelligent video recording and broadcasting system in the student tracking program. The solution use a primary and two secondary cameras to record the video in the classroom, and combine with moving target detection algorithm. This program can handle the situation when more then one student stand up.