Vehicle Speed Evaluation and Prediction Model from DVR Frame Using Feature Detection
In this study, the techniques of digital image analysis, features detection, features correlation and hidden Markov model are employed to develop a vehicle speed emulation and prediction system. The proposed system, all the data source of this study is from a special design vehicular digital video recorder device that includes well defined realtime driving state data such as speed, time, and concurrent frame image. The data of DVR are analyzed to set up the speed evaluation and prediction model. The proposed approaches can closely match the vehicle speed and its concurrent video frame. The results of this study can also provide other vehicular computer vision techniques to reduce the processing time.
vehicular digital video recorder feature detection hidden Markov model
Chun-Nan Lin Chien-Chuan Lin Ming-Shi Wang
Department of Engineering Science National Cheng Kung University Tainan, Taiwan
国际会议
海口
英文
98-102
2011-07-15（万方平台首次上网日期，不代表论文的发表时间）