ForestSemantic: Datasets for Forest Semantic Learning from Close-Range Sensing

ForestSemantic provide detailed annotation of typical objects (e.g., grounds, under-story, trees) and different parts of a tree (e.g., stem, branches, and foliage) in forest scenes by various data types (e.g., imagery, and point clouds).

ForestSemantic includes multiple datasets:

• ForestSemantic_Branch: 3D Branch-level study from Point Clouds
• ForestSemantic_ITD_3D: 3D Tree-level Study by Point Clouds
• ForestSemantic_ITD_2D: Individual Tree Crown Study by Imagery
• ForestSemantic_ITC_Change_2D: Individual Tree Crown Change by Imagery