Train wheel profile segmentation with 2-D minimum cross entropy method based on particle swarm optimization

Train wheel profile image segmentation is a key step of on-line visual inspection system for train wheel dimensions. The 2-D minimum cross entropy thresholding method not only considers gray-level distribution, but also takes advantage of the spatial gray-level distribution, it often gets ideal segmentation results when the image´s signal noise ratio is low. However, its time-consuming computation is often an obstacle in real time application systems. In this paper, the image thresholding method with 2-D minimum cross entropy thresholding method (MCET)based on a new optimization algorithm, namely, the particle swarm optimization (PSO) is presented to deal with train wheel profile image segmentation. The experimental results show that the proposed method can get ideal segmentation results with less computation cost, as illustrated by the portions given in this document.