This paper presents the development and field testing of RoboBoat, an unmanned surface vehicle designed to acquire three‐dimensional (3D) point cloud representations of narrow flooded galleries in caves, mines, and other similar environments where conventional surveying and mapping procedures cannot be applied. We developed techniques tailored to these scenarios, including a contour‐following reactive navigation method and an iterative closest point‐based simultaneous localization and mapping method. Custom metrics have been defined to iteratively validate and improve the 3D model of the passageways. The successful field tests in natural caves and mining galleries demonstrate the accuracy of the 3D representation, making it suitable for digital preservation of large‐volume cultural heritage sites.

中文翻译：

---

RoboBoat：用于对部分被淹没的地下场地进行 3D 测绘的机器人船

本文介绍了 RoboBoat 的开发和现场测试，这是一种无人驾驶地面车辆，旨在获取洞穴、矿井和其他无法应用传统测绘程序的类似环境中狭窄的淹没坑道的三维 (3D) 点云表示。我们开发了针对这些场景的技术，包括轮廓跟踪反应导航方法和基于迭代最近点的同步定位和建图方法。定义了自定义指标来迭代验证和改进通道的 3D 模型。在天然洞穴和采矿坑道中的成功现场测试证明了 3D 表示的准确性，使其适合对大量文化遗产进行数字保存。