Improving the accuracy of DTM and DSM

I asked this question as a follow up question in another thread. Since I did not hear any response there, I am creating this as a new post. Sorry for any duplication.

What are the things that need to be considered to have good DTM and DSM maps from the pix4D?

We have been collecting visual images from our farms at different times in the growing season to see changes in crop heights. For this, we generated DTM and DSM maps along with their Orthomosaics. We used AgRGB template of the pix4D. We came up with pretty odd DTM/DSM maps. The reason is this: when the ground is bare, the DSM values are in the range of 281 - 322. This is close to the existing DEM data for this field. But, for the same field, it came to 280 - 309 when the ground are covered with tall plants. It is kind of odd to see lower DSM values in the later part of the growing season instead of observing higher values.

Is this something other people are encountering as well? Or do I need to select another template for this purpose? Thanks.

Hi Sami,

 

The differences you obtain may be also due to geolocation errors. Do you use GCP’s to align both projects before comparing them?

Which drone do you use to fly your camera?

 

Cheers,

Ina