多摄像头移动识别技术的研究 |
| |
引用本文: | 伊聪聪,吴斌,张红英等.多摄像头移动识别技术的研究[J].工业控制计算机,2014(4):111-113,115. |
| |
作者姓名: | 伊聪聪 吴斌 张红英等 |
| |
作者单位: | 西南科技大学信息工程学院,四川绵阳621010 |
| |
摘 要: | 目前,单摄像头移动对象识别的相关技术已比较成熟,但仍然存在场景切换、目标遮挡等问题,为了解决上述问题及达到更好的识别效果,提出了多摄像头移动对象识别方法,将多视频数据融合技术细化为数据层、目标跟踪层及决策层三个层次的信息融合:数据层通过融合多摄像头移动目标的大小、方向等参数实现了移动目标的准确描述;目标跟踪识别层采用粗集理论和BP神经网络结合的方法,增强了识别系统的稳定性;决策层通过D-S证据理论来融合多个摄像头节点的决策结果,提高了目标识别的准确度。
|
关 键 词: | 移动对象识别 多视频数据融合 粗集神经网络 D-S证据理论 |
Research on Multi-camera Mobile Recognition Technology |
| |
Abstract: | The technology of moving target recognition based on a single camera is much developed nowadays,but there are still many difficult problems.For example,scene change,object occlusion and so on.In order to solve the above problems and achieve better recognition effect,multi-camera moving object recognition method is proposed in this paper.Multi-video data fusion technology is refined into three levels of information fusion including data level fusion,target-tracking level fusion and decision-making level fusion. |
| |
Keywords: | moving target recognition multi-video data fusion rough neural networks D-S evidence theory |
本文献已被 CNKI 维普 等数据库收录! |
|