This paper presents an approach of global sparse matching algorithm based on Delaunay Triangle theory to obtain
reliable matching result of detected feature points between images. Our approach can solve the matching problem in the
case of images captured under a certain range of scaling, rotation and translation, together with image affine distortion,
addition of noise, and change in illumination. Considering that it is hard to obtain a high percentage of correct matched
point pairs, we present this kind of Delaunay Triangle based matching method to improve the accuracy and exactness of
matching result. First, corner detection algorithms are proposed to obtain accurate location of feature points. Then the
relation of the feature points is constructed according to the Delaunay Triangle Theory in the form of a triangle net.
Therefore, the feature points matching problem is transformed into a node angle and length vector matching problem in
the triangle net. During the matching procession, the gray information of the image has not been used. The experimental
results show that the method could achieve accurate matching results with a better percentage of correction.