基于宽孔径双频多波束声呐的水下地形3D重构

Three-dimensional Reconstruction of the Underwater Terrain Based on the Wide-aperture Dual-frequency Multibeam Sonar

针对2D成像声呐数据的水下目标3D重构问题,提出了基于声呐视点球面目标成像原理的3D重构方法。首先,通过引入具有空间变化核的盲反卷积公式,解决了宽孔径声呐数据3D重构的先验不适定问新题;其次,通过关联声呐成像过程中的多个视点,准确恢复了宽孔径声呐数据的高度角信息,该方法适用于主流的商业ROV和AUV,可使用标准声呐传感器实现在线3D重构;最后,将该方法应用于丰满老坝库区采集的双频多波束2D声呐数据,结果表明,重构的3D水下地形与单点测深仪的测绘结果基本一致。

In terms of the three-dimensional(3D)reconstruction of underwater objects from three-dimensional(2D)imaging sonar data,this paper proposes a novel 3D reconstruction method based on the imaging principle of spherical targets from sonar viewpoints.Firstly,by introducing a spatially varying blind deconvolution formula,the a priori ill-posed problem of 3D reconstruction of the wide-aperture sonar data is resolved.Secondly,by correlating multiple viewpoints in the sonar imaging process,the elevation angle information of the wide-aperture sonar data can be recovered accurately.This method is applicable to mainstream commercial ROVs and AUVs,and can utilize standard sonar sensors to achieve online 3D reconstruction.Finally,the method is applied to the dual-frequency multibeam 2D sonar data collected in the reservoir of old Fengman Dam.The results show that the reconstructed 3D underwater terrain is basically consistent with the mapping results of the single-point bathymeter.