A Novel Algorithm for Image Data Segmentation

Abstract

With the rapid development of spatial information technology and the increasingly artificial intelligence knowledge, MAS plays a more and more important role in conducting image segmentation.Considering the shortcomings of current segmentation method,a new algorithm based on MAS theory is proposed, It combines spectral and shape information in region merging.employer a number of agents to control the merging procedure in different regions and make the global merging control more optimal by utilizing the advantages of MAS,The results show that the algorithm is very effective for image segmentation both in urban and mountainous areas.