| 摘要 |
Video motion vector clusters corresponding to a moving vehicle are identified and tracked across multiple frames of captured video. Movement of the motion vector clusters is translated from pixels per second to real speed (e.g. kilometres or miles per hour) to determine whether the vehicle is speeding. Estimated speed data is added to the video stream data as metadata, and video segments of candidate speeding vehicles are stored or transmitted for subsequent automated or manual review. Automated low-complexity video-based vehicle speed estimation is thereby provided. The system may form part of a multi-layer speed enforcement system, in which deeper layers provide enhanced accuracy on selected candidate (speeding) vehicles. |
