三角剖分有限元法在声呐图像重构中的应用  

Research on the Reconstruction of Sidescan Sonar Data Based on Triangulation of Finite Element Method

作　　者：罗进华[1] 杨修伟[2] 朱凌霄[2] 朱培民[2] 

机构地区：[1]中海油田服务股份有限公司,天津300451 [2]中国地质大学(武汉)地球物理与空间信息学院,湖北武汉430074

出　　处：《海洋科学进展》2015年第3期324-333,共10页Advances in Marine Science

基　　金：中海油田服务股份有限公司科研项目--AUV调查数据处理解释系统开发(E-23132019)

摘　　要：侧扫声呐原始数据经几何校正后会造成采样点分布不规则的现象,需要通过重采样获得规则网格数据,以满足海底地貌全覆盖显示的需求。重采样间隔与插值算法对重采样图像的质量具有直接影响。本文分析了侧扫声呐数据采样点的分布特点,在此基础上提出了基于扭曲网格三角剖分有限元插值技术的侧扫声呐图像重构方法。与多种插值方法重采样结果的对比实验表明,该方法不仅提高了计算效率,而且能很好地保留原始数据的信息。Sidescan sonar（SSS）data become irregular after geometric corrections.Resampling is needed in order to get regular grid data,which is needed to get a full coverage SSS image.The interpolation algorithm of resampling and the resampling interval have a direct impact on the quality of the resampled image.This paper analyzed the feature of the SSS data,and proposed a method which was triangulation of distorted meshes of the finite element interpolation method for the SSS data reconstruction.Test shows that this method not only has a higher computational efficiency over other common methods,but also contains more structural information of the original data.

关 键 词：侧扫声呐 扭曲网格 重采样 几何校正 有限元插值 

分 类 号：P756[天文地球—海洋科学]