摘要
在自动视频监控系统中,阴影是影响目标检测和跟踪的主要因素之一。为了去除阴影的干扰,提出一种区域梯度、颜色和轮廓多特征融合的方法。改进轮廓计算方法,巧妙利用轮廓分割区域,合理使用轮廓填充区域,这些改进使本方法更具鲁棒性和更广适用范围。实验结果表明,该方法在不同场景下可以有效地去除阴影。
Shadow is one of the major problems in objects detection and tracking for automated video surveillance systems. In this paper, a method fused regional gradients, color and contour feature is proposed to remove the interference of shadows. The calcula- tion method of contour is improved, separated the region skillfully using contour and rationally utilized the contour-filled region. These improvements make the method be more robust and have broader scope. The experimental results show that the method can effectively remove shadows under different scenarios.
出处
《电视技术》
北大核心
2015年第3期112-114,共3页
Video Engineering
基金
国家科技支撑计划项目(2012BAH20B01)
国家科技支撑计划项目(2014BAK11B02)
广西自然科学基金项目(2012GXNSFAA053232
2013GXNSFAA019326)
广西高校科学技术研究项目(2013YB092)
关键词
阴影去除
特征融合
目标检测
视频监控
shadows removal
feature fusion
objects detection
video surveillance
