基于SFS混合反射模型的侧扫声呐图像三维重建  

3D reconstruction of side-scan sonar images based on the SFS mixed reflection model

作　　者：袁明新[1,2] 张亮 王以龙 刘维 YUAN Mingxin;ZHANG Liang;WANG Yilong;LIU Wei(School of Mechanical Engineering,Jiangsu University of Science and Technology,Zhenjiang 212000,China;Zhangjiagang Industrial Technology Research Institute,Jiangsu University of Science and Technology,Zhangjiagang 215600,China;T-SEA Marine Technology Co.,Ltd.,Zhangjiagang 215600,China)

机构地区：[1]江苏科技大学机械工程学院,江苏镇江212000 [2]张家港江苏科技大学产业技术研究院,江苏张家港215600 [3]中科探海(苏州)海洋科技有限责任公司,江苏张家港215600

出　　处：《海洋工程》2025年第2期187-195,共9页The Ocean Engineering

基　　金：工信部高技术船舶项目([2019]360);张家港市产业链创新产品攻关计划项目(ZKC2206);张家港市产学研预研资金项目(ZKHBZ2303)。

摘　　要：基于明暗恢复形状(shape from shading,简称SFS)方法进行海底侧扫声呐图像三维重建时,会因为方法中反射模型与海底表面不相符而降低重建精度。为此文章在分析大陆架海底表面特点及侧扫声呐工作机理的基础上,提出了基于SFS混合反射模型的侧扫声呐图像海底地形重建方法。首先在保留Lambert漫反射模型的基础上引入Blinn-Phong镜面反射模型,并基于侧扫声呐图像粗糙度进行漫反射系数、镜面反射系数和镜面反射指数的自适应设计,从而形成适合复杂海底表面的混合反射模型;然后线性化处理辐射照度方程,并采用牛顿-拉夫逊方法来获得海底表面高程;最后根据海底点设计反演高程约束系数来约束表面高程,进而完成侧扫声呐图像的三维重建。试验测试结果表明,相较于其他三维重建方法,文中方法的平均绝对误差值VMAE平均降低了32.18%、相关系数值VCC平均提高了29.62%、信噪比值VSNR平均提升了27.23%,有效实现了侧扫声呐图像三维重建。When performing 3D reconstruction of seabed side-scan sonar images using the shape from shading(SFS)method,the mismatch between the reflection model and the seabed surface characteristics often reduces reconstruction accuracy.To address this issue,this study proposes a 3D reconstruction method for seabed topography based on an SFS hybrid reflection model,following an analysis of continental shelf seabed surface features and side-scan sonar operational mechanisms.First,the Lambert diffuse reflection model is retained,and the Blinn-Phong specular reflection model is introduced.An adaptive design for the diffuse reflection coefficient,specular reflection coefficient,and specular reflection exponent is developed based on the roughness of the side-scan sonar image,resulting in a mixed reflection model suitable for complex seabed surfaces.Then,the radiance equation is linearized and solved using the Newton-Raphson method to estimate seabed surface elevation.Finally,the inversion elevation constraint coefficients are designed for seabed points to refine the surface elevation,completing the 3D reconstruction of side-scan sonar images.Experimental results demonstrate that compared to other 3D reconstruction methods,the proposed approach reduces the mean absolute error VMAE by an average of 32.18%,increases the correlation coefficient VCC by an average of 29.62%,and improves the signal-to-noise ratio VSNR by an average of 27.23%,effectively achieving high-accuracy 3D reconstruction of side-scan sonar images.

关 键 词：侧扫声呐图像 明暗恢复形状方法 混合反射模型 三维重建 

分 类 号：P737.2[天文地球—海洋地质]