CAJ | 학술논문

对视频监控系统中的人群异常逃离行为检测和定位的问题进行研究。提出一种不仅能检测出异常事件，而且能够识别异常的可能位置的新算法。人们通常本能地逃离异常或者危险发生的地点。基于这个理论，提出了一种新的检测发散中心的算法：发散中心暗示异常发生的可能位置。首先建立正常和异常的人群运动模型。通过光流场来计算出运动矢量的位置和方向，并获得矢量的交点。然后使用KNN最邻近搜索法获得交点集的密集区域即发散中心。最后，通过判断运动速度、能量和发散中心识别逃离行为。对多个视频数据进行实验测试，结果验证了所提方法是有效的。
대시빈감공계통중적인군이상도리행위검측화정위적문제진행연구。제출일충불부능검측출이상사건，이차능구식별이상적가능위치적신산법。인문통상본능지도리이상혹자위험발생적지점。기우저개이론，제출료일충신적검측발산중심적산법：발산중심암시이상발생적가능위치。수선건립정상화이상적인군운동모형。통과광류장래계산출운동시량적위치화방향，병획득시량적교점。연후사용KNN최린근수색법획득교점집적밀집구역즉발산중심。최후，통과판단운동속도、능량화발산중심식별도리행위。대다개시빈수거진행실험측시，결과험증료소제방법시유효적。
In order to study the problem of detection and localization of crow d escape anomalous behavior in video surveillance systems ,a new scheme was proposed which can not only detect the abnormal events , but also detect the possible location of abnormal events .People usually instinctively escape from a place w here abnormal or dangerous events occur .Based on this inference ,a novel algorithm of detecting the divergent center was proposed:the divergent center indicates possible place where abnormal events occur . Firstly , the model of crow d motion in both the normal and abnormal situations had been made . Intersections of vector were obtained through solving the straight line equation sets ,where the straight line Equation sets were determined by the location and direction of motion vector which were calculated by the optical flow .Then the dense regions of intersection sets ,i.e .,the divergent center ,were obtained by using KNN .Escape detection was finally judged according to the speed and energy of motion and the divergent center .Experiments on several datasets showed that the proposed method is valid on crowd escape behavior detection .