摘要
随着视频采集技术和大规模数据存储技术的快速发展,大型摄像机网络逐渐部署于公共场所。传统的人工监控技术一直难以应对摄像机网络产生的大量视频,而智能视频监控技术已成为解决这一问题的主要途径。行人重新识别作为智能视频监控领域的一个新兴课题,引起了研究人员的广泛关注。行人识别指在现有图像或视频序列中识别目标行人,这些图像或视频序列可能通过非重叠摄像机的视野给定行人图像。一般来说,行人识别图像来自不同的摄像头,其面临着视角、态度、规模和照明变化带来的挑战。目前,如何准确、高效识别行人体重是课题的主要研究方向。
With the rapid development of video capture technology and large-scale data storage technology, large-scale camera networks are gradually deployed in public places. Traditional manual surveillance technology has been difficult to cope with a large number of videos generated by the camera network, and intelligent video surveillance technology has become the main way to solve this problem. Pedestrian recognition, as a new topic in the field of intelligent video surveillance, has attracted extensive attention of researchers. Pedestrian recognition refers to the recognition of target pedestrians in existing image or video sequences, which may be given a pedestrian image through the field of view of non-overlapping cameras. Generally speaking, pedestrian recognition images come from different cameras, which are facing challenges brought by changes in perspective, attitude, scale and lighting. At present, how to accurately and efficiently identify pedestrian weight is the main research direction of the subject.
引文
[1]巫细波,杨再高.智慧城市理念与未来城市发展[J].城市发展研究,2010,17(11):56-60.
[2]刘治红,骆云志.智能视频监控技术及其在安防领域的应用[J].兵工自动化,2009,28(4):75-78.
[3]曾明勇,吴泽民,田畅,等.基于外观统计特征融合的人体目标再识别[J].电子与信息学报,2014,36(8):1844-1851.