Research of Fast FCM Vehicle Image Segmenting Algorithm Based on Space Constraint

Abstract

Vehicle identification and traffic accident detection plays an important role in the Intelligent Transportation System. Vehicle image segmentation is the key technological foundation for further identification and detection processing. This article improves the inadequacy of Fuzzy C-Means (FCM) clustering algorithm by proposing the Spatial Constrained FCM (SCFCM) algorithm. Firstly, the each pixel's membership degree is corrected according to its field pixels', eliminating the impact of noise on the accuracy of FCM clustering. Secondly, a new searching algorithm based on Gaussian model single-peak judgment is proposed to obtain the optimal number of clusters. After that, initial membership matrix creation algorithm is used to reduce iteration times. The performance of the experiments shows that method is effective.