Parts Classification based on PSO-BP

Abstract

Aiming at the problems of low accuracy and low efficiency of the traditional parts classification method, and the strong nonlinear mapping relationship between the feature parameters of parts and the types of parts, this paper proposes a parts classification and recognition method based on particle swarm optimization (pso) and BP neural network (bp). Firstly, the acquired part image is preprocessed to extract image features such as affine invariant moment, circularity and rectangularity. The image feature parameters and the corresponding part categories constitute the data set through PSO-BP training, and then different kinds of parts are identified by the trained classifier. Compared with the BP neural network classification, the particle swarm optimization BP neural network classification is not easy to fall into the local minimum solution, and the classification and recognition accuracy is higher. The experimental results show that this method has higher classification and recognition accuracy in the process of parts sorting.