三维计算机视觉及在场景分割解释中应用  

3D computer vision and its application in scene segmentation and interpretation

作　　者：杜谊[1] 陈新[2] 吴芳[2] 

机构地区：[1]福建医科大学网络中心,福建福州350004 [2]福州大学信息科学与技术学院,福建福州350002

出　　处：《福州大学学报（自然科学版）》2001年第z1期72-74,共1页Journal of Fuzhou University(Natural Science Edition)

基　　金：福建省科技三项费用资助项目 (K2 0 0 13 )

摘　　要：介绍一种实用的三维立体计算机视觉方法 ,以自动汽车的简单立体视觉系统为例 ,利用一些与立体视觉相对应的关键特征 ,结合多目标并行图象信息和色彩信息进行场景描述 ,同时使用场景中有关物体的先验知识和已知模型等 ,基于这些描述、知识和模型 ,大多数的场景目标可以根据预处理所获得的信息进行有效的分割解释处理 .This paper describes applications of 3D computer vision in companies and our approach to a more flexible 3D computer vision. In order to overcome shortcomings of the current stereo vision,a reliable stereo method is proposed which makes use of the reliability of features in determining correspondence. A rich description of a scene is obtained by integrating range information and color information. Interpretation of a scene requires the use of knowledge or models of objects. Based on the knowledge,most effective process is triggered at each stage depending on what kind of information is obtained up to the stage.

关 键 词：三维计算机视觉 多目并行图象 场景分割 场景解释 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]