A fusion method for TOF point cloud and digital photograph

摘要

Image registration is to find the transformation parameters between images, which is in order to eliminate or reduce the image nonlinear or linear deformation caused by the imaging equipment or imaging conditions. In this paper, an image registration model based on the relationship of geometric projection imaging is established. First of all, proposed a calibration method based on planar pattern, shot in fixed color image plane calibration template on the color checkerboard pattern under different angles and amplitude image, improved Harris corner extraction algorithm, using Levenberg-Marquardt algorithm to obtain the TOF and CCD camera parameters, and the like the relationship between the plane pose estimation of rotation matrix and translation vector between the two camera coordinate system, finally realize the integration of TOF point cloud and digital photograph using this algorithm. The experimental results show that this method can improve the efficiency of calibration and improve the registration accuracy.