Research on Artificial Target Image Matching

Image matching is one of most important problems in close-range photogrammetry. In order to improve the speed of image matching, many artificial targets are placed in measurement scene. There are two kinds of artificial targets, encoded target and non-coded target. How to match the target images is discussed in detail in the paper. An image matching method for encoded targets is introduced, in which several geometry invariants is used. An image matching method for non-coded target is presented, which make full use of the image matching results for en-coded targets, and in which epipolar geometry constraint is used. The three dimensional reconstruction is used in the method, which make the image matching for non-coded targets more fast and. exact A software is developed based on the method. An experiment has been done, and the results show the method is correct and effective.