Image matching algorithm based on optimal sampling RANSAC

The paper proposes an image matching algorithm based on optimal sampling random sample consensus (OSRANSAC) to handle the shortcoming (eg. high computational costs and poor system performance) for visual positioning system. Scale-invariant feature transform (SIFT) algorithm is used to extract feature points and generate feature descriptors, Meanwhile, the fast approximate nearest neighbor algorithm is adopted to complete rough matching of feature points. In order to reduce the number of iterations, we employ in cross-combination resampling algorithm to calculate model parameters by the new selected sample set, and apply pre-test method to choose the best model parameters. Simulation results confirm the superiority of the proposed methods in comparison with classic ones in terms of the accuracy rate and real-time performance.