Dear All and Esri,
I would like to suggest a new deep learning model to be built to extract 3D trees from a 3D Mesh of whole municipality (city/town etc.).
I have tried Tree Point Classification, and I would say it is doing a fantastic job identifying tree point cloud.
However, not all cities having luxury collecting ortho imagery and LiDAR data at the same time, and deep learning model for Tree Detection or Tree Segmentation are not working well as the Tree Point Classification model, which make sense.
Or maybe there is a way to convert 3D Mesh to 3D Point Cloud, and I can still use Tree Point Classification model to extract tree points?
My goal is to generate tree canopy coverage data in the end utilizing 3D Mesh with Tree Point Classification deep learning model! since my organization will only need to collect ortho imagery + 3D mesh every two years to gain accurate tree canopy data without collecting LiDAR at the same time.
Any guidance will be appreciated! Hope this idea will be supported 🙂
Many thanks!
