The Algorithm Of Micro Moving Target Extraction And Background Replacement In The Video Sequences

With the rapid development of computer technology and communication technology, technological innovation of the communication between the users and the operators is undergoing tremendous change by the video sequences. The technology mainly involved the micro moving target detection and extraction, in which the key factor is to improve the accuracy of moving target detectionn and to overcome the interference of the environmrnt. Recently, a lot of moving target detection algorithms have been put forward, but the detection algorithm for micro moving targets were relatively less. So the study of the micro moving target detection and background replacement algorithm which in the video sequences has important theoretical and practical value.The micro moving target extraction and background replacement algorithm in the video sequences was studied in this thesis, which aimed to improve the accuracy and real-time of the micro moving target extraction. Firstly, the thesis came up with the algorithm that combined the Otsu and the skin detection for the micro moving target extraction and background replacement. Secondly the algorithm of the micro moving target extraction was immproved based on LBP background modeling, then a agent video system was designed, and achieved verification and performance analysis. The main research work as follows:1. The algorithm of image segmentation and moving object detection is studied, and the algorithm of partially micro moving target extraction is analyzed and summarized.2. In the thesis, the micro moving target extraction algorithms in the video sequences were studied depthly. The thesis proposed a micro moving target extraction algorithm which combined the Otsu and the skin color detection. This algorithm included local image enhancement, micro moving target extraction, holes filling, background replacement to achieve the micro moving target extraction accurately and complete the background replacement.3. The traditional moving target detection could not extra the micro moving target from the video sequence, because of the complex background and the shadows. In order to improve the accuracy of the micro moving target extraction from the complex background, the thesis studied the problem of shadows depthly, the micro moving target extraction based on the LBP background modeling was improved, and the shadows removal algorithm was proposed. The experimental results showed that the improved background model based on LBP effectively improve the accuracy of the micro moving target extraction.4. Based on the micro moving target extraction and background replacement algorithm, the thesis designed and implemented an agent video system. The system could not only process the input video sequences in real time, and eextract micro moving target and finish the background replacement, but also transmit the new video sequences to the phones.Finally, the thesis summaried the research work and pointed out the future research direction, which extended our thoughts in the future.