The Research on Behavioral Status Judgment Method Based on Skeleton Modeling and Vector Processing

Abstract

Fast, simple and reliable method for judging the behavioral status of video object is the public research base of many application fields, including public safety monitoring, national defense, aeronautics, sports and so on. This paper focuses on the fast parametric skeleton modeling and simple process of vectorization. It proposes an improved thinning algorithm for fast skeleton modeling and then locates the key points by a new scanning and marketing method, ultimately, according to the coordinate of the key points, it constructs a matrix of pose estimation and formulates the behavioral status judgment as a problem of calculating the angle between vectors. This method not only resolves the problems of large volumes of data and high complexity of algorithm in the traditional approach, but also lays the foundation for the extended researches and applications of multi-field intelligent video behavior analysis.