Improved Image Matching Method Between SAR and Optical Based on Gaussian Vote by Ballot
摘要
Aiming to obtain the high-precision matching position and reduce time and resource requirements, a SAR and optical image matching method based on CSP-DenseNet and Gaussian vote by ballot is proposed. First, in order to detect stable interest points and extract deep feature in SAR image, we use FAST operator and CSP-DenseNet. Then, in order to save time and resource requirements, we use gaussian vote by ballot to determine the high-precision matching position between two images. The simulation results reveal the validity of the algorithm.