Abnormal Behavior Recognition Based on Multi-Information Decision in Power Generation

Abstract

So far most research of human behavior recognition focus on simple individual behavior, such as wave, crouch, jump and bend. This paper will focus on abnormal behavior with objects carrying in power generation. Such as using mobile communication device in main control room, taking helmet off during working and using umbrella in high place. In global environment, there are some almost fixed features for portable objects of workers, such as clothes. So we adopted edge detecting by color tracking to recognize object in worker. The sequence of 3D human behavior data would be expressed by geometric character of skeleton and its angle. Since specific time and space of behavior definition, it is of importance for recognition system using describing information in a comprehensive way. Take account of information complementary perspective, this paper introduced a method of making decision from multi-information by spatial pyramid and fisher score discretion for behavior recognition. And the method was proved to have advantages than some previous algorithms.