The field of 3D plant phenotyping and point cloud analysis is rapidly advancing, with a focus on developing innovative methods for plant species identification, organ segmentation, and point cloud classification. Recent studies have highlighted the potential of deep learning architectures, such as EfficientNetB0 and Point-SkipNet, in achieving high accuracy and efficiency in these tasks. The development of new datasets, such as ModelNet-R, and open-source frameworks, like Plant Segmentation Studio, are also bridging the gap between algorithmic advances and practical deployment. Noteworthy papers in this area include:
- treeX: Unsupervised Tree Instance Segmentation in Dense Forest Point Clouds, which presents a revised version of the treeX algorithm for efficient tree instance segmentation.
- Enhancing 3D Point Cloud Classification with ModelNet-R and Point-SkipNet, which introduces a refined dataset and a lightweight graph-based neural network for high classification accuracy.
- Towards scalable organ level 3D plant segmentation: Bridging the data algorithm computing gap, which provides a comprehensive review of existing 3D plant datasets and deep learning-based methods for point cloud semantic and instance segmentation.