Matching and Recognition Method of Architectural Decoration Elements Based on Hybrid Particle Swarm Optimization Algorithm

Abstract

TImage matching and recognition of architectural decoration elements is the basis of architectural decoration design. The traditional matching and recognition methods have low precision and high complexity, so they are not suitable for the image matching and recognition of architectural decoration elements under complex conditions. Based on the analysis of the basic particle swarm optimization algorithm, the nonlinear asynchronous strategy is adjusted to change the fixed constant mode of the learning factor to balance the local and global search ability of the algorithm in the iterative process. At the same time, the vitality factor was introduced to mutate the inactivated particles to improve the population diversity. The experimental results show that compared with the traditional algorithm, this method improves the image matching and recognition accuracy of architectural decoration elements.