摘要: 为了方便计算机视觉中的图像配准和拼接等工作,该文提出了一种最小化畸变平面校正方法。通过求射影变换权值变化测度之和的极值及Jacobian行列式度量校正前后的局部区域变化,求得最小化重采样效果的校正变换矩阵。建立广义平面校正坐标系,在虚拟坐标系内考虑最小化畸变和最小化重采样效果,减少了旋转造成的重采样畸变,改善了可视性。
关键词:
计算机视觉,
图像校正,
最小化畸变,
图像拼接
Abstract: In order to be in favor of image registration and mosaic in computer vision, a planar rectification which minimizes the deformation is studied. The rectification transform matrix is solved to seek the extremum of sum of projective transform weight change and minimize resampling effect that is weighed by Jacobian determinant. The generalized planar rectification coordinate system is built and normalized hypothetically. Minimizing distortion and resampling effect are calculated in the hypothetically normalized coordinate system. It keeps the similarity before and after transforming, and the visibility is improved.
computer vision; image rectification; minimizing distortion; image mosaic
Key words:
computer vision,
image rectification,
minimizing distortion,
image mosaic
中图分类号:
郭君斌;郭晓松;杨必武;雷 磊. 基于最小化畸变重投影的广义平面校正方法[J]. 计算机工程, 2007, 33(22): 26-28,7.
GUO Jun-bin; GUO Xiao-song; YANG Bi-wu; LEI Lei. Generalized Planar Rectification Method Based on Minimizing Distortion Reprojection[J]. Computer Engineering, 2007, 33(22): 26-28,7.