作　　者： ; ; ; ;

机构地区： 韩山师范学院

出　　处： 《计算机仿真》 2013年第6期389-392,共4页

摘　　要： 研究图像匹配的人流量密度估计,提高估计的准确率。针对在基于人流图像的密度估计过程中,人流存在着突变性、不可预见性、非线性等特点,当瞬时的人流量密度发生较大突变的情况下,例如短时内人流密度大幅增加或减少,人流密度变化会影响估算结果,造成人流密度估计不准确的问题。为了解决上述问题,提出一种采用物联网的人流密度估计方法,通过采集不同时刻的图像人流信息,运用人流图像计算补偿匹配技术,充分考虑图像帧之间的特征联系,对下一帧与前一帧图像的差异人数量特征进行补偿,弥补传统方法的缺陷。实验证明,改进方法能准确估计人流大幅变化时的密度,取得了不错的效果。 Research based on communities density image matching, improve estimated accuracy. Based on the image density estimation in the process, there are people, not the foresight, retinoblastoma occurring nonlinear characteristics, when the density of the instantaneous teaming up in great mutations, such as in a short flow density increased or decreased, and the changes will affect people the estimated results, causing the density flow density esti- mation inaccurate problem. In order to solve this problem, the paper proposes a kind of based on content flow of den- sity estimation method networking, through the acquisition of different time image information, people use people im- age calculation compensation, considering the image matching technology of contact between frames to the next, fea- tures and a frame before images of differences between people compensate the quantity characteristics, make up for the defect of the traditional method. Experiments show that the method can accurately estimate people in a density of change, and achieved good effect.

关 键 词： 人流密度估计 物联网 补偿匹配

领　　域： [自动化与计算机技术] [自动化与计算机技术]