Article Preview
Top1. Introduction
Visual science is an important way to help human individuals perceive the external environment world and obtain psychological information of the external environment. Due to its unique characteristics of external space-time and space, visual environment perception often plays an extremely important leading role in the human body perception processing system of all kinds of people, and the environmental information that can be obtained and obtained from the human body's visual perception system is also more intuitive and rich. With the rapid and in-depth development of mobile Internet and multimedia information technology, the amount of visual information in the real-life world continues to grow rapidly. If directly making a computer itself have the ability to calculate motion perception based on human body vision, then it can directly fill all of these needs. Although individual human beings are the main body of social relations, their behaviors and psychological actions often have specific social meanings (Ding, 2019; Chang 2019). Accurate recognition of various human behaviors can often directly help to better understand the social and psychological behaviors of various people. Therefore, accurate recognition of human actions has developed into one of the most active research topics in the field of visual science in the current computer era.
Automatic tracking and surveillance video automatic surveillance technology based on human-computer vision is a multi-disciplinary academic interdisciplinary research topic that has attracted much attention from all walks of life in recent years. The positioning detection and automatic tracking technology of the facial image of the moving robot under the background of complex information is the current academic hotspot and research difficulty in the development of various fields of computer automatic vision technology research in the world. There are extensive research applications in various fields such as navigation, traffic management, multimedia teaching, target perspective recognition and automatic tracking, and security automatic monitoring. At the same time, it can also be the theoretical basis for various image subsequent advanced image processing technologies such as image target perspective classification, behavior pattern recognition and logical understanding. The main research results of this research topic can not only be widely used in solving various illegal motion captures, but also can be applied to safety automatic monitoring (Wan, 2019; Wei, 2019). And further in-depth research on the automatic vision of the robot, the creation of the map library for automatic positioning and tracking of the image of the mobile robot in the unknown information environment, the question bank and the tracking of the target person are of important research significance. This research is based on the GMM algorithm and the stereo matching algorithm to encode the human motion capture algorithm of the video, which makes the human motion capture algorithm more sensitive.