How do I get accurate DTM from DSM to get tree height?

Question

I have drone imagery which was generated in Agisoft. From the imagery a DEM was generated and exported.

I would like to know how do I get the Digital Terrain Model from this DEM? as I currently have the Digital Surface Model?

The method I have tried doing was creating points at different locations in between the trees(on the ground) and interpolating the points to create a DTM by subtracting the DSM and DTM (DSM-DTM= tree height) from each other.

However the results are not consistent enough to achieve tree heights accurately.

I do have ArcGIS Pro.

Answer 1

I am currently doing a research on how to create a more precise DTM to calculate plant height for plants under 2 meters (needs to be very precise). There are 3 methods to create a DTM:

• Collecting data when there is only ground. This will be your DTM (can be done for crops since they grow seasonal, but no in the case of trees).
• Interpolate points from the ground.
• Use automatic classification. I have use automatic creation of DTM in PIX4DMapper and it has resulted in almost the same accuracy or even better than the DTM from colecting a dataset of only ground. The reason for this is you do not have the overlaping inacuracy between 2 datasets (even when using GCPs or RTK or PPK, there is always some error).

Drone suggestions to improve DTM:

• High side & front overlap (80-90)
• Low GSD (fly lower and also with a high resolution camera)
• Make sure drone is updated and calibrated (IMU, Compass) before each flight
• If using GCPs, make sure there is high image overlap where the GCPs are. If they are in the border, increase the flight mission margin

Answer 2

Just subtract the DEM from the DSM. As you said, DSM - DEM = CHM (DTM). No need to create points, just use the rasters.

You can do this in ArcGIS Pro using the raster calculator: DSM - DEM

https://www.earthdatascience.org/courses/use-data-open-source-python/data-stories/what-is-lidar-data/lidar-chm-dem-dsm/