CAJ | 학술논문

针对传统车辆检索方法中存在准确性和区分度较低的问题，提出了一个基于改进SURF （ speeded up robust features）算法的视频车辆检索方法。在车辆视频关键帧提取的基础上，根据改进SURF算法完成车辆图像的特征提取及匹配，其中包含改进FAST （ features from accelerated segment test）特征点检测、SURF特征向量提取以及最近邻查询方法来进行特征点的匹配；通过计算比较待检索车辆图像与数据库车辆图像的相似度，算法完成图像筛选并反馈检索结果。实验结果表明：针对交通监控视频中待检索车辆，该方法能够较为准确地进行检索并反馈结果。
침대전통차량검색방법중존재준학성화구분도교저적문제，제출료일개기우개진SURF （ speeded up robust features）산법적시빈차량검색방법。재차량시빈관건정제취적기출상，근거개진SURF산법완성차량도상적특정제취급필배，기중포함개진FAST （ features from accelerated segment test）특정점검측、SURF특정향량제취이급최근린사순방법래진행특정점적필배；통과계산비교대검색차량도상여수거고차량도상적상사도，산법완성도상사선병반궤검색결과。실험결과표명：침대교통감공시빈중대검색차량，해방법능구교위준학지진행검색병반궤결과。
To overcome the problem low accuracy and discrimination of traditional vehicle retrieval methods, a new vehicle video retrieval method based on improved SURF algorithm is proposed. On the basis of vehicle video key frame extraction, the improved SURF algorithm is used for extracting and matching of vehicle image features, inclu-ding improved FAST angular point algorithm to extract the image feature points, including SURF algorithm to extract the image feature vector and including nearest neighbor query algorithm to get matching points;through cal-culating and comparing vehicle images to be retrieved and the database of vehicle images in similarity, image filte-ring is completed and the retrieval results are output. The experimental results and their analysis show preliminarily that this method not only can detect the video vehicle, but also can feedback retrieval results fairly accurately.