| Video surveillance is an important means of ensuring safe running of railway. Video surveillance system would capture the real-time video signal of the key regions and transmit to the monitoring center. The operator can easily deal with the abnormal condition by the video image. Traditional video surveillance system has strong dependence on the operator judgment, so that it often can not detect abnormal condition in the monitoring video in time. To let the system automatically detect abnormal condition and accurately alert is a hot spot in the current research of the railway video surveillance. Using video analysis technology to detect and identify moving objects will be a good solution to this problem.Being base on the background of Shenhua Group research project " Shenhua Railway Broadband Wireless system", this paper combined with the requirement of railway operation safety video alert, made a depth research in the key technologies of the railway traffic safety video warning, then brought up the moving target detection algorithm based on Gaussian mixture model and the moving target recognition algorithm based on fuzzy theory. The algorithms work on the video image segmentation and extraction, and it can effectively identify the moving objects in the railway traffic safety scene. The main works are as follows:(1)Analyzed and researched on moving targets detection, compared and validated in the railway video surveillance scene, then put forward the target detection method for railway background.(2)Researched on the moving target fuzzy recognition algorithm for railway background, then the result validated by experiment shows that the algorithm of moving object on the railway scene classification.(3) Design of railway traffic safety video warning system, description of the detailed design of the main functions of the system, the result shows that the system of the identification and automatic warning for the moving objects in the railway traffic safety. |
PDF Full Text Request