Modelling and simulation: An improved RANSAC algorithm based on the relative angle information of samples

Random sample consensus (RANSAC) algorithm is the most widely used one in the field of computer vision. In order to reduce the high complexity of RANSAC, this paper proposes a novel method which can reject samples before calculating the homography matrix. This algorithm can eliminate random samples that may be wrong through calculating the relative angle information of the random samples, and then, use the correct samples for the next step. The algorithm can ensure the accuracy of the premise while greatly reducing the computational complexity. Not only that, the improved algorithm can also be combined with the existing RANSAC extensions to improve the computational efficiency.