测绘学报 ›› 2023, Vol. 52 ›› Issue (10): 1669-1678.doi: 10.11947/j.AGCS.2023.20220579

• 大地测量学与导航 • 上一篇    下一篇

结合不确定度与密度聚类算法的多波束异常值自动滤波算法

王俊森1, 金绍华1, 边刚1, 崔杨1, 龙振宇1,2   

  1. 1. 海军大连舰艇学院军事海洋与测绘系, 辽宁 大连 116018;
    2. 91937部队, 浙江 舟山 316002
  • 收稿日期:2022-10-18 修回日期:2023-04-20 发布日期:2023-10-31
  • 通讯作者: 金绍华 E-mail:jsh_1978@163.com
  • 作者简介:王俊森(1998-),男,硕士生,研究方向为多波束水深测量数据处理。E-mail:beanwjs@163.com
  • 基金资助:
    国家自然科学基金(41876103)

A multi-beam outlier automatic filtering algorithm combining uncertainty and density clustering method

WANG Junsen1, JIN Shaohua1, BIAN Gang1, CUI Yang1, LONG Zhenyu1,2   

  1. 1. Department of Military Oceanography and Hydrography, Dalian Naval Academy, Dalian 116018, China;
    2. Troops 91937, Zhoushan 316002, China
  • Received:2022-10-18 Revised:2023-04-20 Published:2023-10-31
  • Supported by:
    The National Natural Science Foundation of China (No. 41876103)

摘要: 本文在复现CUBE滤波算法的基础上,借鉴其网格节点可吸收水深点选取模型,提出了一种结合不确定度与密度聚类算法的多波束异常值自动滤波算法。本文使用DBSCAN密度聚类算法对水深值加以聚类,使用卡尔曼滤波推估节点水深值,选取具有最小不确定的水深假设作为节点水深值,实现对多波束测深数据异常值的有效清理。实测数据和仿真试验结果表明:CUBE滤波算法不能将连续异常值完全剔除,而本文算法能够较好地去除连续异常值。本文算法流程明晰、参数简单、性能可靠,对数据质量较差的情况下较多异常值也能够进行清理,具有实际的工程应用价值。

关键词: 多波束测深, 异常值自动滤波, CUBE算法, DBSCAN算法, 卡尔曼滤波

Abstract: Based on the reproduction of CUBE filtering algorithm, this paper proposes a multibeam automatic outlier filtering algorithm combining uncertainty and density clustering method in reference of CUBE's assimilation model. In this paper, we use the DBSCAN to cluster the bathymetry values, use Kalman filter to estimate bathymetry values of the node, and select the bathymetry hypothesis with minimum uncertainty as true bathymetry values of the node. The measured data and simulation results show that the CUBE filtering algorithm cannot completely eliminate the continuous outliers, while the algorithm in this paper can clean up the continuous outliers better. Our algorithm is clear, simple and reliable, and can clean up many outliers in the case of poor data quality, which possesses practical engineering application value.

Key words: multi-beam bathymetry, outlier filter, CUBE algorithm, DBSCAN algorithm, Kalman filter

中图分类号: