PTZ摄像机监控中的运动目标检测  

Moving object detection in PTZ camera surveillance

作　　者：裴银祥 周兵[1] 

机构地区：[1]郑州大学信息工程学院,郑州450001

出　　处：《高技术通讯》2014年第2期152-156,共5页Chinese High Technology Letters

基　　金：863计划(2008AA01A315)资助项目

摘　　要：为实现摄像机运动下的安全监控,提出了一种不需要摄像机标定的运动目标检测算法。该算法分两个阶段进行,第一阶段使用图像跟踪技术估算每一帧的摄像机转动量,当转动量累加到阈值时,抓取当前帧作为背景并用累加转动量标记该帧,重复此过程获取所有背景帧序列并建立索引;第二阶段根据当前摄像机的转动量选取最接近的背景帧,利用快速图像配准算法配准当前帧和背景,采用背景差分检测出变化像素和运动对象。实验表明,该算法在普通PC上能够实现实时检测。The moving object detection under free-moving camera was studied, and a new moving target detection algo- rithm without camera calibration was presented to realize the video surveillance with a PTZ camera. The algorithm is divided into two stages. In the first stage, background sequences were snapped and indexed. Camera motion parame- ters were estimated through tracking each background image. Once the camera pan parameter accumulation was up to a threshold, a background frame was captured and indexed with the pan value. In the second stage, the moving object detecting was carried out. Firstly, the closest background frame was chosen with the camera＇ s current pan value, and then, the motion parameters were estimated between the current frame and the background picture based on the image registration algorithm ,which was used as an initial parameter to track the camera and detect moving objects until a new background frame was replaced. The experimental result showed that the proposed algorithm can be applied to real-time moving object detections with common PCs.

关 键 词：运动对象检测 相位相关 图像配准 背景减 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]