Target state recognition of basketball players based on video image detection and FPGA

Abstract In the optimization for target tracking of basketball players, the traditional approach has a major drawback. The function results in a highly non-convex function is defined as a linear combination of spatial and appearance constraints on the target. Very dense graph has been configured to capture the attributes of the target. Video detection is based on sports player game movement, then edge detection based on video running. They were moving the ball position for each player. FPGA configuration is used for edge detection algorithm is fully supported from playbacks. If the system is verified, it shows a video image is an edge detecting system that can detect a highly accurate edge. Learning techniques for the extraction of deep foundation, started by the main features, and significant progress has been made in this respect. The latest progress uses various learning techniques of in-depth and detailed investigation in target detection results. Several topics are covered, including gradient direction, histogram, a single detector and a double shot, datasets, indicators, acceleration, and the current state of art probe. Some important applications in the object detection region include a detailed discussion of the detection and detection of a pedestrian population.