Spatiotemporal Matching Cost Function Based on Differential Evolutionary Algorithm for Random Speckle 3D Reconstruction

Random speckle structured light can increase the texture information of object surface, so it is added in binocular stereo vision system to solve matching ambiguity problem caused by surface with repetitive pattern or no texture. To improve reconstruction quality, many current researches utilize multiple patterns for projection and use methods based on spatiotemporal correlation. This paper presents a novel random 3D scheme, which are used weighted-fusion-based cost function (STMCF) proposed find corresponding points image pairs. Furthermore, parameter optimization method differential evolutionary (DE) algorithm designed automatically determining values all parameters included STMCF. In this method, since there suitable training data ground truth, we explore strategy where passive dataset truth as then apply learned value Various experimental results verify that our scheme realize accurate high-quality efficiently STMCF exhibits superior performance terms accuracy, computation time quality than state-of-the-art