The processing results, dataset and original codes of Coral reef PtCloud segmentation model (based on proposed underwater camera systems)

# Data IntroductionThe proposed camera system includes three operation modes: Surface Mode, Horizontal Mode, and Curved Mode. Surface Mode: Provides 3D reconstructions (point cloud), DEM, and orthophoto maps of the seabed (covering a line of 70m in length). Horizontal Mode: Provides 3D reconstructions (point cloud and mesh) and orthophoto maps of a single transect line. Curved Mode: Offers 3D reconstructions (mesh) of various targets, including artificial coral reefs, coral reefs with snails, and coral reefs with starfish. All mesh results are saved in .FBX format, and point cloud results are saved in .TXT format. # Code IntroductionThe backbone of our point cloud segmentation model is the KPConv model.If you encounter any issues during code deployment, please refer to the original KPConv repository (Original Code: https://github.com/HuguesTHOMAS/KPConv).We provide our modified code (customized for our task) along with the complete point cloud dataset.# 数据集简介 本次提出的摄像系统包含三种工作模式：海底表层模式（Surface Mode）、水平勘测模式（Horizontal Mode）以及曲面扫描模式（Curved Mode）。 海底表层模式（Surface Mode）：可生成海底的三维重建成果（点云）、数字高程模型（DEM，Digital Elevation Model）以及正射影像图，覆盖长度为70米的测线范围。 水平勘测模式（Horizontal Mode）：可生成单条测线的三维重建成果（点云与网格模型）以及正射影像图。 曲面扫描模式（Curved Mode）：可对多种目标生成三维网格重建结果，涵盖人工珊瑚礁、附着螺类的珊瑚礁以及附着海星的珊瑚礁。 所有网格模型结果均以.FBX格式存储，点云结果则以.TXT格式保存。 # 代码说明 本研究的点云分割模型的主干网络为KPConv模型。若在代码部署过程中遇到任何问题，请参考原始KPConv代码仓库（原始代码：https://github.com/HuguesTHOMAS/KPConv）。我们提供了针对本任务定制的修改版代码，以及完整的点云数据集。