海洋测绘2025，Vol.45Issue(5)：1-5,5.DOI:10.3969/j.issn.1671-3044.2025.05.001
基于剖分与异构加速的快速测深精度评估
Rapid evaluation of bathymetric accuracy based on partitioning and heterogeneous acceleration
摘要
Abstract
To rapidly and efficiently evaluate the accuracy of massive bathymetric point cloud data,based on the traditional method of constructing seabed terrain surfaces using triangulated irregular networks,this study proposes an improved scheme for calculating master-check verification discrepancy values of massive bathymetric data that integrates the adaptive convex hull partitioning algorithm with a CPU-GPU heterogeneous computing architecture.Specifically,the convex hull algorithm is used to automatically delineate the coverage area of check lines,replacing the manual method for coverage area delineation;heterogeneous parallel computing is employed to accelerate triangulation and the calculation of master-check verification discrepancy values.Experimental results show that this scheme achieves a 4-10 fold improvement in processing speed with measured multibeam data in the South China Sea,which can efficiently process over 100 million bathymetric point cloud data points,and the calculation accuracy of master-check verification is consistent with that of traditional methods.关键词
多波束测深/主检比对不符值/精度评估/凸包算法/CPU-GPU异构Key words
multibeam bathymetry/master-check verification discrepancy values/accuracy evaluation/convex hull algorithm/CPU-GPU heterogeneity分类
天文与地球科学引用本文复制引用
秦玉峰,欧阳永忠,洪黎丹,何西,陆长锴..基于剖分与异构加速的快速测深精度评估[J].海洋测绘,2025,45(5):1-5,5.基金项目
国家自然科学基金(42174013) （42174013）
自然资源部海洋环境探测技术与应用重点实验室开放基金(MESTA-2023-A002) （MESTA-2023-A002）
自然资源部南海局2024年度科技发展基金(240103) （240103）
2024福建省科技重大专项(科教联合)(2024HZ026009) （科教联合）
福建理工大学引进科研教学团队资金(GY-24015). （GY-24015）
