运动目标检测与跟踪算法的改进与实现  

Improvement and realization of the moving object detection and tracking

作　　者：杨冬冬[1] 常丹华[1] 韩夏[2] 刘宇[1] 

机构地区：[1]燕山大学信息科学与工程学院,河北秦皇岛066004 [2]南京邮电大学电子科学与工程学院,江苏南京210003

出　　处：《激光与红外》2010年第2期205-209,共5页Laser & Infrared

摘　　要：提出了一种对视频图像中运动目标检测并进行实时跟踪的新方法。该方法利用基于背景建模的背景差分与改进的带时间戳的运动历史图像(tMHI)的目标分割相结合的算法对运动目标进行检测,在获取到视频流中的运动目标轮廓后使用基于tMHI的运动梯度算法来实现运动目标的跟踪。实验结果表明,该方法能够对指定区域内的目标进行有效识别和准确跟踪,并且弥补了运动目标暂时性停止时无法检测出来的不足。A new method for detection and tracking of real-time moving object in video image was purposed. This method uses background difference based background modeling and developed timed motion history image to detect the moving object. The algorithm of motion gradients based timed motion history image is applied to track the moving object after getting the silhouette of moving object. The result of test shows that this method can detect and track the moving object in the given region effectively. The system can remedy the deficiency of fail detection when the moving object stops temporarily.

关 键 词：带时间戳的运动历史图像 背景建模 运动目标检测 运动目标跟踪 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]