回溯搜索优化算法辅助的多阈值图像分割  

Backtracking search optimization algorithm assisted multilevel threshold for image segmentation

作　　者：尹雨山 王李进[1,2] 尹义龙[1,3] 王冰清[1] 赵文婷[1] 徐云龙[1] 

机构地区：[1]山东大学计算机科学与技术学院,山东济南250101 [2]福建农林大学计算机与信息学院,福建福州350002 [3]山东财经大学计算机科学与技术学院,山东济南250014

出　　处：《智能系统学报》2015年第1期68-74,共7页CAAI Transactions on Intelligent Systems

基　　金：国家自然科学基金-广东联合基金重点资助项目(U1201258);山东省自然科学杰出青年基金资助项目(JQ201316)

摘　　要：阈值法是一种简单且有效的图像分割技术。然而阈值求解的计算量随阈值的增加而呈指数级别增长,这给多阈值图像分割带来巨大挑战。为了克服计算量过大问题,视多阈值分割模型为优化问题,分别将Otsu法和Kapur法作为目标函数,采用回溯搜索优化算法求解目标函数,实现多阈值图像分割。将提出的多阈值分割算法应用于自然图像分割,并与其他算法比较,实验结果说明基于回溯搜索优化算法的多阈值图像分割技术是可行的,而且具有较好的分割效果。The threshold method is a simple and effective image segmentation technique. However,the amount of calculation for solving threshold appears to be exponential amplification with the increase of threshold. This results in a huge challenge for multi-threshold image segmentation. This paper utilizes Otsu and Kapur methods as the target function in order to deal with image segmentation.In this paper,image segmentation is considered as an optimization problem whose objective function is formulated according to Otsu and Kapur methods,respectively. The backtracking search optimization algorithm is used to solve these two objective functions and to realize multi-threshold image segmentation. The proposed approach is applied to nature image segmentation and compared to other algorithms. The results showed that the multi- threshold image segmentation technique on the basis of backtracking search optimization algorithm is feasible and the segmentation effect is satisfactory

关 键 词：阈值法 回溯搜索优化算法 图像分割 OTSU Kapur PSNR 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]