HOG内嵌模板匹配的车门夹人检测算法

An algorithm for detecting train door persons based on HOG inline template matching

我国新建地铁多采用屏蔽门设计,但屏蔽门和列车车门之间存在间隙,容易卡、夹乘客造成严重的安全事故.针对这一问题,提出基于视频帧图像的车门夹人检测技术.首先,采用HOG算子描述候车乘客的图像特征,并提取类乘客目标的凸显特征;然后提出一种动静目标分离算法,将乘客从背景中分割出来,并实现对乘客的有效实时跟踪.最后,利用现场模拟数据开展实验,结果表明本文提出的检测算法可以实现对车门夹人的有效检测.

Many new subways in China are designed with screen doors, but there is a gap between the screen doors and the train doors, which makes it easy to clip passengers and cause serious safety accidents. To solve this problem, this paper proposes a technology to prevent people from being caught by the door based on video frame images. First, the HOG operator is used to describe the image characteristics of the waiting passengers, and the salient features similar to passengers are extracted. Then a dynamic and static target separation algorithm is proposed to separate the passengers from the background and achieve effective real-time tracking of passengers. Finally, experiments are performed using on-site simulation data. The results show that the proposed detection algorithm can achieve effective detection of the doors.