旋转图像序列的整合  

CONFORMITY OF ROTATING IMAGE SEQUENCE

作　　者：吴福朝[1] 于洪川[1] 袁波[1] 韦穗 

机构地区：[1]安徽大学人工智能研究所,合肥230039

出　　处：《模式识别与人工智能》1999年第3期249-254,共6页Pattern Recognition and Artificial Intelligence

摘　　要：本文讨论图像整合问题,给出由摄像机旋转运动产生序列图像的整合方法。该方法有以下优点:整合图像仍满足透视几何关系,这保证了整合图像的真实性;整合参数的变化,对应于在不同视点下的整合,即可得到不同视点下的整合图像;此外,该方法不需要事先知道摄像机的内、外参数。这些优点说明本文的方法具有广泛的应用价值。In this paper, the conformity of rotating image sequence is discussed and the method of conformity is presented. Its merits include the following, (l)The conformity image accords with persective geometry. This one ensures that the conformity image is true. (2)The conformity coefficient corresponds to view. When the conformity coefficient is changed, the conformity image corresponding to different view can be obtain. (3)In practice, the camera needn't to be calibrated. So this method is very practical.

关 键 词：匹配矩阵 图像整合 整合参数 图像序列 图像编码 

分 类 号：TN919.8[电子电信—通信与信息系统] TP391.41[电子电信—信息与通信工程]