联合多特征提取与内部优化的海陆分离  

Sea-land segmentation algorithm based on multi-feature extraction and interior optimization

作　　者：程红[1] 刘思彤[1] 孙文邦[1] 

机构地区：[1]空军航空大学,吉林长春130022

出　　处：《激光与红外》2015年第1期99-103,共5页Laser & Infrared

基　　金：全军军事类研究生资助课题(No.2013JY514)

摘　　要：目前大多数海陆分离方法依靠先验信息或仅利用灰度特征对图像进行处理,对导致的误分割或大量孤立区域无法较好处理,造成后续处理工作难度大。针对这一问题,提出一种联合多特征提取与分离结果优化的海陆分离方法。针对不同的陆地类型提取不同的特征进行分割,对得到的结果采用面积与纹理相结合的方式进行优化,进而得到最终的海陆分离结果。实验结果表明,本文方法与常用的海陆分离方法相比能够快速完整地实现海陆分离并将陆地屏蔽,有利于后续处理工作的进行。Most sea-land segmentation algorithms resulted in wrong segmentations or the generation of many isolated areas by using only gray feature or relying on prior information,which seriously affect the subsequent processing.Ai-ming at this problem,a sea-land segmentation algorithm based on multi-feature extraction and interior optimization is proposed.Firstly,the suitable features are extracted to segment for the different land types.Then the obtained results are optimized by the combination of area and texture.Finally,land is separated from the sea.The results show that the proposed algorithm can segment the sea and land exactly and effectively,and it can also shield the land,which pro-vides convenience for subsequent processes.

关 键 词：海陆分离 多特征提取 内部区域优化 陆地屏蔽 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]