Research And Implementation For Target Tracking And Privacy Protection Based On Video Stream

Along with the rapid development and progress of computer communication and multi-media techniques, Computer-Supported Cooperative Work (CSCW) is now becoming a very active research field in computer science and technology, and the theory of collaborative system has been widely used in fields such as video conferencing, multi-functional perception, multi-media CAI, cooperative discussion and so on. Privacy protection in the video images is an important security technique in remote collaborative systems. So it is very necessary to develop a real-time and convenient protection method used in video streams of collaborative system.First, we address the importance of the privacy protection in collaborative system and introduce the correlated key techniques such as image processing, target detection, moving object tracking and so on. We present the current research status of the related fields in detail and their further development trend.Then, we carefully describe the whole process of analysis, design and completion of the approach of target tracking and privacy protection in real-time environment, primarily focusing on object detection in color images, moving target tracking in video sequences and privacy protection. I will particularly discuss the algorithm of region segmentation based on projection and the improved tracking method based on moving prediction.When discussing the target tracking and the privacy protection in video sequences, we introduce all related techniques, which includes color modeling of the object, modeling method, noises eliminating approaches, region segmentation, region merging, feature matching, moving forecast and tracking and so on.Finally, applying all these techniques mentioned above, we have designed and developed a privacy protection system in real-time video environment. The functions and implementation of the system, as well as its applications, are described in the end of this thesis.