Abstrakt

Biomechanical model-based sprint athlete articulation point identification method research

Chunhong Wu, Li Wang

Video human sports analysis is an important research direction in computer research field, main content is getting human movement information from single or multiple video camera shooting video sequence, and further carry on three-dimensional reconstruction, with modern human movement capturing and behaviors understanding’s research development, it has higher and higher requests on the research[1]. The paper grasps sports movement process human articulation point automatic recognition scientific algorithm as key points to make discussion, in the hope of building basis for more authentic and scientific three-dimensional reconstruction. The paper implements flow research by human articulation point automatic recognition software, and states image processing latter period human body model principle of model that reconstructs according to proportion. Apply video camera measuring apparatus and scientific measuring method, provides basis for effective and scientific data collection. Research result defines human body 15 links and 21 articulation points, and provides each articulation point defining mathematical model, establishes special status discriminant model, provides visible lateral articulation point entering into human contour line, invisible lateral articulation point sheltering and hand definite impacts on hip joint processing methods.