摘要
为了满足实时性的要求,在运动目标检测与跟踪领域引入感兴趣区域(Region of Interest,ROI)的概念,提出了基于矩形扩张的ROI区域标记算法.通过与传统算法比较发现,矩形张算法具有较高的效率和准确度.实现结果表明:矩形扩张算法能够快速准确地对二值图像中的运动目标区域进行标记,实现运动目标与复杂背景的分离,进而提高运动目标跟踪的效率.
This paper introduces the idea of ROI （ Region of Interest） to moving target detection and tracking research,and proposes the ROI labelling algorithm based on rectangular expansion. The algorithm can label moving targets in binary image quickly and accurately,thus separates moving targets from complex background effectively and improves the efficiency of moving target tracking. Comparison between this ROI labelling algorithm and traditional dynamic clustering algorithm shows that the former can label moving targets with less time complexity and higher efficiency.
出处
《南京信息工程大学学报（自然科学版）》
CAS
2010年第6期573-576,共4页
Journal of Nanjing University of Information Science & Technology（Natural Science Edition）
基金
南京林业大学科研发展基金(X08-300-1)
