A Novel and Effective Cooperative RANSAC Image Matching Method Using Geometry Histogram-Based Constructed Reduced Correspondence Set

The success of many computer vision and pattern recognition applications depends on matching local features two or more images. Because the initial correspondence set—i.e., set feature pairs—is often contaminated by mismatches, removing mismatches is a necessary task prior to image matching. In this paper, we first propose fast geometry histogram-based (GH-based) mismatch removal strategy construct reduced Creduced,GH from Cini. Next, an effective cooperative random sample consensus (COOSAC) method for remote sensing COOSAC consists RANSAC, called RANSACini working Cini, tiny RANSACtiny,GH randomly selected subset Creduced,GH. RANSACtiny,GH, iterative area constraint-based sampling proposed estimate model solution Ctiny,GH until specified confidence level reached, then utilizes estimated calculate inlier rate repeats above cooperation between reporting resultant For convenience, our GH-COOSAC method. Based several testing datasets, thorough experimental results demonstrate that achieves lower computational cost higher accuracy benefits when compared with state-of-the-art methods.