We are currently completing a research project into the use of drones for forestry. We will investigate whether we can provide critical information for wood volume calculations. This will involve automating tree diameter calculations and tree heights. Tree heights will be done by subtracting a DTM from a DSM to create a Canopy Height Model (CHM).
It will be determined if photogrammetry in these cases can provide a good terrain model since there could be difficulty penetrating the canopy to do so. Again, airborne LiDAR (Technology) would be the best technology here and it is something we are moving towards; in the interim we intend to fully explore the applicability of photogrammetry in this case.
The next investigation will be how to automate tree diameter calculations, which will likely involve using computer vision and training an object detection model.
Another piece of information critical to foresters is tree health. We will be using a multispectral sensor (using the DJI Phantom 4 Multispectral drone) to capture information beyond the potion of light visible to humans (RGB) with a focus on the NIR band to compute an NDVI index to yield a measure of photosynthetic activity. We will also be looking at computer vision to bring this down to a tree level and also for automated tree species recognition to identify invasive species.