基于CV模型和形状信息的光学遥感舰船目标分割方法  

Ship target segmentation in optical remote sensing images based on CV model and shape information

作　　者：张辉[1] 杜春[1] 孙浩[1] 计科峰[1] 

机构地区：[1]国防科学技术大学电子科学与工程学院,长沙410073

出　　处：《东南大学学报（自然科学版）》2013年第A01期168-173,共6页Journal of Southeast University：Natural Science Edition

摘　　要：为了克服光学遥感舰船目标分割中的光照、视角变化和阴影问题,提出一种基于CV模型和形状信息的光学舰船目标分割方法.该方法首先利用核主成分分析来提取舰船目标的形状信息,然后将构造的形状能量泛函与CV模型相结合,得到新的舰船目标分割模型.最后,提出了新的Heaviside函数,以确保水平集演化过程中的数值稳定性.选用来自不同传感器的光学遥感舰船目标图像进行实验.实验结果表明,提出的方法在光学遥感舰船目标的分割性能上优于传统的CV模型.Ship target segmentation in optical remote sensing images meets a serious challenge,owing to the variations of light conditions,visual angle and the shadow problem.To address these difficulties,a ship segmentation method based on Chan-Vese（CV） model and shape information is presented.First,kernel principle component analysis（KPCA） is used to extract shape information of optical remote sensing ship targets.The obtained shape energy term and the CV energy term are then combined into a new ship segmentation model.To ensure the numerical stability of symbol function of the proposed model,a new Heaviside function is also presented in the evolvement process of level set.The ship images from different optical remote sensors are selected for segmentation to evaluate the performance of the proposed method.Experimental results show that the method can contribute better segmenting effect than original CV model.

关 键 词：分割 CV模型 KPCA Heaviside函数 

分 类 号：U459.2[建筑科学—桥梁与隧道工程]