Abstract
An algorithm for image matching of multi-sensor and multi-temporal satellite images is developed. The method is based on the SIFT feature detector proposed by Lowe in (Lowe, 1999). First, SIFT feature points are detected independently in two images (reference and sensed image). The features detected are invariant to image rotations, translations, scaling and also to changes in illumination, brightness and 3-dimensional viewpoint. Afterwards, each feature of the reference image is matched with one in the sensed image if, and only if, the distance between them multiplied by a threshold is shorter than the distances between the point and all the other points in the sensed image. Then, the matched features are used to compute the parameters of the homography that transforms the coordinate system of the sensed image to the coordinate system of the reference image. The Delaunay triangulations of each feature set for each image are computed. The isomorphism of the Delaunay triangulations is determined to guarantee the quality of the image matching. The algorithm is implemented in Matlab and tested on World-View 2, SPOT6 and TerraSAR-X image patches.
Highlights
Most of the older and recent researches (Lowe, 2004, Lowe, 1999, Feng et al, 2008, Yang and Kurita, 2013, Harris and Stephens, 1988, Moravec, 1981, Shi and Tomasi, 1994, Zhao and Ngo, 2013, Harris, 1993) on image matching and registration are based on the concept of detecting feature points in the reference image and matching them to the corresponding feature points in the other image
In order to solve this problem, local interest points with, as much as possible, unique geometrical, topological and spectral characteristics have to be detected. These points should be highly distinctive in the sense that they can be identified successfully against a large database of other points. This uniqueness of feature points is necessary in image matching because in most of the cases in real life, images taken at different dates or/and from different sensors are at the same time rotated, translated and different in scale and illumination
We have presented a novel quality control technique based on Delaunay triangulation isomorphism to assess Scale Invariant Feature Transform (SIFT)-based image matching
Summary
Most of the older and recent researches (Lowe, 2004, Lowe, 1999, Feng et al, 2008, Yang and Kurita, 2013, Harris and Stephens, 1988, Moravec, 1981, Shi and Tomasi, 1994, Zhao and Ngo, 2013, Harris, 1993) on image matching and registration are based on the concept of detecting feature points in the reference image and matching them to the corresponding feature points in the other image. In order to solve this problem, local interest points with, as much as possible, unique geometrical, topological and spectral characteristics have to be detected These points should be highly distinctive in the sense that they can be identified successfully against a large database of other points. The ground-breaking work of Lowe in 1999 (Lowe, 1999) extended the local feature based previous approaches even more by proposing a scale invariant method, the prominent Scale Invariant Feature Transform (SIFT) method Even though it is not a transform, it is called transform in the sense of transforming image data into scale-invariant coordinates(Lowe, 2004).
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
