| 基于空时置信关系的运动检测方法 |
| |
| 引用本文： | 范志辉,陈曦,王辉. 基于空时置信关系的运动检测方法[J]. 计算机应用研究, 2018, 35(7) |
| |
| 作者姓名： | 范志辉 陈曦 王辉 |
| |
| 作者单位： | 河南科技大学 洛阳 471023,河南科技大学 洛阳 471023,河南科技大学 洛阳 471023 |
| |
| 基金项目： | 国家自然基金(61502364),河南省重点科技项目基金 (132102210246),河南省产学研项目(142107000022),赛尔下一代网络创新项目(NGII20150311),河南科技大学博士科研启动基金。 |
| |
| 摘 要： | 为了提高运动目标检测的准确度和精度,提出一种基于空时置信关系的运动检测方法。该方法利用快速核密度估计对图像像素点与其邻域像素点的空时关系进行建模,并根据样本值的离散度为背景模型分配对应的权重,最后依据像素值的背景隶属度权重均值,判断当前像素点属于运动前景还是背景。实验结果表明该方法的运动检测性能优于主流代表性算法。
|
| 关 键 词： | 背景建模 目标检测 运动检测 运动分割 空时置信关系 |
| 收稿时间： | 2017-02-06 |
| 修稿时间： | 2018-05-26 |
Motion Detection Based on Spatio-temporal Confidence Relationship |
| |
| FAN Zhihui,CHEN Xi and WANG Hui. Motion Detection Based on Spatio-temporal Confidence Relationship[J]. Application Research of Computers, 2018, 35(7) |
| |
| Authors: | FAN Zhihui CHEN Xi WANG Hui |
| |
| Affiliation: | Henan University of Science and Technology,, |
| |
| Abstract: | A novel scheme based on spatio-temporal confidence relationship is designed to improve the performance of motion detection. This method employs fast kernel density estimation to model the relationship of a pixel and its neighbors, designates appropriate weight to the pdf(probability distribution function) of environmental points according to the disperson degree of their sample values, and discriminates a target pixel belonging to background or foreground in a similar voting method by fusing the judging results of the environmental point set. Experiments on typical surveillance scenes verify that the proposed algorithm performs better than the state-of-the-art methods. |
| |
| Keywords: | Background modeling Object detection Motion detection Motion segmentation Spatio-temporal confidence relationship |
|
| 点击此处可从《计算机应用研究》浏览原始摘要信息 |
|
点击此处可从《计算机应用研究》下载全文 |
