水下坝面多源设备检测数据集

本数据集来源于国家重点研发计划项目“水电站大坝缺陷智能识别诊断与精细检测设备及技术研究”（课题编号：2022YFB4703404），面向水利工程安全检测与水下结构病害智能识别研究，基于水下机器人搭载的高清光学相机、多波束声呐与三维激光扫描仪协同观测产生。数据来源于2024年6月至2025年10月对三峡、白鹤滩等水库大坝的定期巡检，主要记录了坝面缺陷的高分辨率视频、声呐图像及高精度三维点云等多模态观测值，原始数据量600余MB。采集遵循“声呐初检、光学详查、激光精测”的流程：先以声呐进行大范围扫描并录屏，再操控机器人抵近拍摄清晰的光学缺陷视频，最后在重点区域进行三维激光扫描与实时建模。数据集按“工程-传感器类型-数据文件”的层级组织，包含剥落、裂缝、露筋等典型缺陷的多源对应数据，并通过严格的设备校验、过程监控与标准化存储管理确保质量，为水下机器人多传感器融合与自动化检测算法研发提供了真实、全面的基准数据资源。

This dataset is derived from the National Key R&D Program of China project "Intelligent Identification, Diagnosis and Precision Detection Equipment and Technology for Hydropower Dam Defects" (Grant No. 2022YFB4703404). It is developed for research on hydraulic engineering safety inspection and intelligent identification of underwater structural diseases, and is generated via collaborative observations using an underwater robot equipped with high-definition optical cameras, multi-beam sonars, and 3D laser scanners. The data was collected during regular inspections of reservoir dams including the Three Gorges Dam and Baihetan Dam from June 2024 to October 2025. It mainly records multi-modal observation data such as high-resolution videos, sonar images, and high-precision 3D point clouds of dam surface defects, with the original data volume exceeding 600 MB. The data collection follows the workflow of "initial sonar inspection, detailed optical inspection, precision laser measurement": first, perform large-range scanning and screen recording using sonar, then maneuver the robot to approach the target for capturing clear optical defect videos, and finally conduct 3D laser scanning and real-time modeling in key areas. The dataset is organized in a hierarchical structure of "Project - Sensor Type - Data File", and contains multi-source corresponding data for typical defects such as spalling, cracks, and exposed rebar. Its quality is guaranteed through strict equipment calibration, process monitoring, and standardized storage management, providing authentic and comprehensive benchmark data resources for the development of underwater robot multi-sensor fusion and automated inspection algorithms.

搜集汇总

背景与挑战

背景概述

该数据集是一个面向水利工程安全检测的水下坝面多源设备检测数据集，基于水下机器人搭载的高清光学相机、多波束声呐与三维激光扫描仪协同观测产生，包含高分辨率视频、声呐图像和高精度三维点云等多模态数据，主要记录剥落、裂缝、露筋等典型坝面缺陷，为多传感器融合与自动化检测算法研发提供真实基准资源。