一种视频监控中的夜间车辆检测方法

A Night Vehicle Detection Method For Video Surveillance

车辆检测是车辆跟踪和车牌识别的前提,在电子警察和视频监控系统中有重要作用。针对目前视频监控中对夜间车辆检测方法较少且检出率不高的问题,提出了一种夜间车辆检测方法。首先选取夜间车辆样本作为训练样本,然后利用AdaBoost算法训练级联检测器,对视频监控中的夜间车辆进行检测。结果表明,本文提出的夜间车辆检测算法能够有效的检测夜间车辆,检测速度快,可实时检测。

Vehicle detection is the premise of vehicle tracking and license plate recognition, whichplays an important role in the electronic police and the video surveillance system. In view of the cur-rent video surveillance on the night vehicle detection method is less and the detection rate is nothigh, put forward a night Vehicle detection method. First choose the night vehicle samples as train-ing samples, and then use AdaBoost algorithm training Cascade detector, and finally the use oftrained cascade detectors in the video surveillance of night vehicles to detect. The results show thatthe night vehicle detection algorithm can effectively detect the night vehicle, the detection speed isfast and can be detected in real time.