Research on abnormal behavior target tracking algorithm in airport intelligent video surveillance

Abstract

With the rapid development of China's civil aviation industry, the airport is facing increasing pressure on security. In this paper, the target tracking algorithm in Intelligent Video Surveillance (IVS) is studied. It aims to provide ideas and reference for the development and implementation of high performance intelligent video surveillance system. The main contents of this paper are as follows: Aiming at the problem of tracking failure caused by occlusion, deformation and illumination changes, this paper proposes a target tracking algorithm that combines the apparent features and depth characteristics. Firstly, the CNN network is trained by a large number of pedestrian databases, and then the depth characteristics of the target area are extracted by trained CNN network. At the same time, the color histogram of the target area in HSV space is calculated, and the depth feature and color feature are combined to get the whole feature. Finally, a number of hypothetical states are estimated under the framework of particle filter, the optimal position of the target is obtained, the tracking result is obtained, and the template is updated. Finally, the resampling is carried out according to the degeneration of the particle. Experiments show that the tracking algorithm has good tracking robustness. Finally, the target tracking system is designed and simulated on the Matlab platform. The validity and practicability of the algorithm are verified.