The segmentation of wear particles in ferrograph images based on an improved ant colony algorithm

Ferrography is a notably useful means to determine the wear condition of machines. Before attempting to extract the feature parameters of wear particles for identification and analysis, it is necessary to separate wear particles in ferrograph images. Hence, wear particle segmentation is a critical first step for intelligent ferrography based on computer image analysis. This paper presents a new method for the segmentation of wear particles by combining watershed and an improved ant colony clustering algorithm. The experimental results have demonstrated the possibility of achieving accurate segmentation of wear particles, including large abnormal wear particles and deposited chains.