Locating the position and orientation of an image by the automatic image registration technique

This report explains the use of an automatic image registration technique to locate the positions and orientations of unidentified camera-captured images. The techniques for solving the shifting- and rotating-object problems are investigated. In this research, one standard straight-sample image and various unidentified camera-captured images are also stored in the system. The algorithm developed in this study can detect the positions and orientations of unidentified images and move them toward straight-sample images. By this approach, both types of images can be aligned to the same position and orientation. After a straight-sample image is overlapped with an unidentified image, the latter is subtracted from the former. By this approach, one can identify the accuracy rate of the developed algorithm. Under normal circumstances, when photographs of the objects are taken, although the positions of the objects are slightly shifted or rotated in other directions, the developed algorithms can correctly identify the position and orientation of the object.