Research Of Video Image Algorithm Of Personnel Fluidity Statistics

In public places like railway stations, shopping malls, plazas and public channel, personnel fluidity can be miscellaneous and complicated, often involves the issue of public security and traffic jams. Using video image detection technology to monitor and identify the stream of people and some moving will improve the public management efficiency significantly. This thesis mainly research the core technologies of motion target detection, pedestrians identification and pedestrians motion tracking for counting the number of pedestrians in the video surveillance area based on the theory of digital image processing. The main contents are as follows:This thesis using image preprocessing to remove noise from the image and improve the contrast of the image to improve the quality of the image because of poor quality raw video image. Preparing for follow-up image processingIn moving target detection, to improve the frame difference method and the background subtraction method that set dynamic threshold to adaptive changing environment based on the analysis of common moving target detection algorithm, and combine background subtraction and frame difference of two algorithms to use of complementarity of themselves. The experimental results show that the algorithm can overcome the noise, detection point and hole image feature extraction of the phenomenon;In pedestrians identification, the algorithm is the use of the overall characteristics of the information of the human body such as area, perimeter, ratio of height to width and the ratio of the difference between the feature parameters and other characteristic parameters of a moving object and the moving body in the mark, so that the number of mobile personnel can be counted based on the analysis of common pedestrian recognition algorithm. The test shows that the proposed algorithm can effectively identify the human object and the number of people from the video images of the mobile man and objects;In pedestrians tracking, the algorithm take the advantage of the characteristic that the track of pedestrian in video surveillance image moves smoothly within the time interval of a few frames, and then using the pedestrian area and centroid parameters to get the movement trajectory of human body based on the analysis of common pedestrian tracking algorithm. The test shows that the proposed algorithm can effectively track of multi moving human targets.The research of this thesis can realize the dynamic statistics of the video surveillance image, which has a strong theoretical and practical value for the development of the field of intelligent video surveillance.