An improved algorithm for panoramic image matching

Aim to the disadvantage of traditional panoramic image feature extraction and matching algorithm, such as huge data processing and low time efficiency, an improved panoramic image feature extraction and matching algorithm base on SURF and improved KD tree search is proposed. Firstly, SURF is used to extract feature point and generate feature description, secondly, the KD tree search method with high efficiency is used to match feature point, and an improved KD tree search method is proposed to avoid wrong matching. Simulation and experiment results in case of translation and rotation indicate that the proposed algorithm owns better efficiency, accuracy and robust property.