The construction method of measurable aerial panorama based on panoramic image and multi-view oblique images matching

Abstract

This paper presents a method to construct the measurable aerial panorama based on panoramic image and multi-view oblique images matching. It's a new way that the visual expression of three-dimensional geographical environment is provided by stitching multi-view UAV images into aerial panorama images, which is widely used in smart city, smart travel, safety emergency and many other fields. The problem that aerial panorama images can't be located and measured is attribute to the limitation of panorama image stitching algorithm and the traditional geographic reference model. This paper proposes a method for panoramic images and multi-view oblique images matching, which can solve this problem admirably. The main steps contain projection, matching, and back projection. The method starts from generating the image by projecting the aerial panoramic image according to the degree. Then, a set of seed points are exacted by matching between the projected image and the corresponding multi-view images. Finally, the seed points in the projected image back projection on the aerial panoramic image. Experiments demonstrate the accuracy of our approach, and the result is satisfactory.