Perspective invariant binary feature descriptor based image matching algorithm

Current local feature based image matching algorithms are usually less robust to image perspective transformation. Aiming to solve this problem, a new perspective invariant binary code (PIBC) based image matching algorithm is proposed. Firstly, FAST corners are detected on the pyramid images, those corners with non-maximum Harris corner response value and the edge points are further eliminated. And then, by simulating the perspective transformations of images taken from different viewpoints, a single FAST corner is described with binary descriptors under different viewpoint transformations, which makes the descriptor could describe the identical feature point on different perspective transform images. Experimental results show its robustness to image perspective transformation, while its complexity is similar with SURF. ©, 2015, Editorial Board of Journal on Communications. All right reserved.