Video Behavior Detection and Recognition Based on Background Contrast

Video behavior detection is a popular research area in video understanding and has been widely and importantly used in unmanned operations, defense industry, and human-computer interaction. In this design, a system that can detect and recognize human behavior in video is designed. The video is first sub-framed, the background frame images are extracted and the images are pre-processed, for the subsequent extracted frame images. Differences are made with the background frames respectively, then morphological processing is carried out to remove the interference impurities, enhance the target region, get the human body contour, and detect the tracking of the moving target. Finally, the behavior detection and recognition are done to complete the final system design. Through the design and verification of the above, this design is characterized by high accuracy.