基于图像角特征的三维运动估计

利用图像角特征来研究计算机视觉领域中的运动估计问题。首先通过引入一个新的坐标系，由单幅图像求得直角三维结构的解析解，其次根据两幅图像中相对于同一个角各条边的关系，求得物体的旋转运动，最后通过求解一个简单的线性方程组得到物体的相对平移运动，并证明了由两幅视图的一对角匹配对和两对点匹配对就能唯一地确定物体的运动。这种基于直角特征的运动估计方法还可推广到非直角的一般情形，实验结果表明该算法是正确有效的。

3D Motion Estimation from Image Corners

The motion estimation from image corners was investigated. By introducing a new coordinate system, the three dimensional structure of an orthogonal corner can be easily recovered from a single image frame. According to the relationship of the angle edge lines between two image frames, the rotation motion can be computed, and the translation is estimated up to a scale factor by solving linear equations. It is concluded that one corner and two points correspondences over two views are sufficient to uniquely determine the motion. The algorithm is extended to the general corners which are not orthogonal. The experimental results show the correctness and efficiency of the algorithm.