论文部分内容阅读
提出了一种新的多摄像机视觉监控系统的信息融合方法 .信息融合在两个阶段进行 .首先 ,根据相互独立的Cartesian参考坐标系统 (设置在地平面上 ) ,对各个摄像机进行标定 .然后 ,把所有的坐标系变换到一个坐标系统中 .在视觉监控应用中 ,因为摄像机自定标和视觉数据配准技术将使监控设施安置变得更加容易 ,从而可以为公共场合发展更加适用的视觉监控工具 .在解决监控数据的不完整性和不确定性方面 ,机器学习方法具有很好的效果
A new information fusion method based on multi-camera vision monitoring system is presented.The information fusion is carried out in two stages.Firstly, each camera is calibrated according to the independent Cartesian reference coordinate system (set on the ground plane) Transforming all coordinate systems into a single coordinate system In visual surveillance applications, as camera self-calibration and visual data registration techniques make monitoring facilities easier to deploy, more suitable visual surveillance can be developed for public use Tools.Machine learning methods have a great effect in resolving the incompleteness and uncertainty of monitoring data