Overlap Analysis of the Images from Unmanned Aerial Vehicles

Compared with traditional images, high-resolution images obtained by unmanned aerial vehicles (UAV) are superior in many aspects such as low cost, convenience and so on. However, the attitude of unmanned aerial vehicles cannot be determined accurately, which will lead to the decrease of accuracy and efficiency of automatic matching, so image preprocessing should be done to guarantee enough degree of overlap for image mosaicking and mapping. There are some ascendant characteristics of SIFT operator such as invariant to scale, rotation, and lightness, and these characteristics are very useful in increasing the matching accuracy of UAV images. In this paper, the overlap analysis of UAV images has been processed based on the SIFT algorithm, and the steps of overlap analysis are described, and the way of treating wrong matching points is explained. Finally, satisfactory results are obtained.