Population Density Of The Research And Implementation Of Automatic Statistics System

With the promoting of the Green City, the construction and application of video surveillance is more extensive. How to effectively manage for the crowd in an important place in order to ensure the safety of the population has become an urgent problem. The traditional method is to manually monitor the video screen, which will cause fatigue and greatly reduce the efficiency, and can not real-timely get the crowded conditions of the population, and then to take certain warning measures. So it is Urgent to study the statistics algorithm of population density based on intelligent video analysis by computer, instead of manually judgment.In this paper, images and video features of high definition surveillance video have been investigated. The population density is analyzed through statistical regression based on the body edge and the human motion area characteristics. The paper also studied image pre-processing, self-adaptive parameters and sub-window weighted technology.The video algorithm analysis uses advanced image processing and choices different pattern recognition. The algorithm is as shown:automatic mode, day mode, night mode, the festival mode, learning mode, where the first four modes uses interframe difference algorithm and edge extraction algorithm. This extract movement area and pedestrian edge through the morphological processing of images and draw crowds to area ratio of total effective area to calculate the fitting parameters, and then given the population density and its rating. The paper is based on Innovation video analysis algorithms and developed behavior analysis algorithms and software demonstrations of key areas. This is supplementary analysis of the system as the population density.The main contents are as follows:1. The paper reviewed the development of the existing video surveillance systems and the application of video analysis technology in video surveillance. The several existing crowd density statistic algorithm is discussed specially and the advantages and disadvantages are analyzed.2. The motion area and object edge detection based crowd density statistic algorithm is proposed in the paper. The implement step is given, and how to apply in different surveillance scene is discussed in detail.3. SVM based crowd density statistic algorithm is studied in the paper. The algorithm is tested and discussed based on OpenCV.4. The architecture of the high-definition video surveillance system has been researched. The crowd density statistics software is developed based on the surveillance system and is applied in the Nanjing Confucius Temple Square, and the validity and advanced of the proposed algorithm is proved.