Xu Xiaojun,Li Jianhua,Wang Sun'an,Guo Yonghong
(School of Mechanical Engineering,Xi'an Jiaotong University,Xi'an 710049,China)
Abstract:Based on the principle of 2-demensional entropy£¬an improved algorithm for image segmentation is put forward,where the spatial correlation and grey scale are considered£¬and the target edgy is detected to kept down more edgy informations.The threshold is searched by optimum family genetic algorithm£¨OFGA£© which has the ability to change the search space and population size.Compared with the other similar segmentation algorithm,the newly proposed one facilitates accelerating the computing rate,improving the accuracy and preventing the prematurity.To illustrate the algorithm validity,the Lena image whose size is 256¡Á256 pixels is segmented 100 times.The results show that the average time reaches to 1.593 7 s,the average evolution number approaches to 2.503 7£¬and the edgy informations are kept down perfectly£®
Keywords:2-demensional entropy;image segmentation;genetic algorithm;edgy detection;grey