Representative rice plant point clouds and CBAM-UNet model for the Metatlas V1 workflow

This record contains eight representative 3D point clouds of four rice cultivars at the heading and maturity stages, with one point cloud provided for each cultivar at each stage. The point clouds were acquired using the Metatlas V1 multi-view imaging platform and reconstructed using COLMAP and OpenMVS. A Python script is provided for extracting plant height and canopy width from the point clouds. Version 2.2 adds a minimal inference script and expanded documentation for image preprocessing and binary-mask generation. Version 2.1 expanded the representative point-cloud dataset by adding V2 and V4 point clouds at both the heading and maturity stages. Version 2.0 added the CBAM-UNet model implementation and the corresponding trained checkpoint used for rice plant image segmentation in this study. The representative point clouds and trained checkpoint are released under the Creative Commons Attribution 4.0 International license. The phenotypic trait extraction code is released under the MIT License. The CBAM-UNet source code is distributed under the GNU General Public License v3.0 only. Further details are provided in the accompanying README and LICENSES files.