Hello all,
I recently bought a Matrice 300 RTK pro with the L1 lidar sensor.  I flew a site last weekend PPK setting up a base station over surveyed point with known coordinates.  I processed using DJI terra and the .las file appears to be about 80' off vertically and about 10' off horizontally.  I was able to check the results because the site had previously been flown with my Matrice 600 and compiled photogrammetrically using that same base station point.  The problem would either be at time of flight or post processing but I'm not sure where to look.  Any suggestions?



I also posted this question in the post processing section.

Is the reference surface orthometric?

LV_Forestry Posted at 4-11 08:34
Is the reference surface orthometric?

HI,
No, the reference surface is a topographic map with contours and planimetric features compiled from stereo imagery.

That's not my question.  

There are two ways to measure altitude.

The first according to the international reference used by GNSS systems, which is more commonly called "ellipsoidal height".  It is a geometric shape that represents the average curvature of the earth.  It is this reference that indicates the altitude of DJI drone data.  

The second way is to rely on mean sea level.  It is a system that is used locally because the earth is not round at all, its shape is very random.  This is the Orthometric height.  

It is therefore necessary to start by identifying which coordinate system is used in your reference raster in order to apply a transformation.  If you don't know it, you can apply the EGM96 correction which should get closer to the truth.  It is an international geoid model.

If you can't, send me your raster, I'll tell you what it's made of. And therefore which raster transformation is to be applied.

HI,
I asked the head of our surveying dept, he said we are using mean sea level orthometric height.

fans1c07f6ff Posted at 4-11 09:37
HI,
I asked the head of our surveying dept, he said we are using mean sea level orthometric height.

So here is where the discrepancy comes from.  

You can ask him what reference geoid he uses, then all you have to do is shift the point cloud from GRS80 provided by L1, to your reference system.

Agisoft Metashape does it very well.  Otherwise with Global Mapper + Geographic Calculator.  There are surely others, these are the two that I use.

fans1c07f6ff Posted at 4-11 09:37
HI,
I asked the head of our surveying dept, he said we are using mean sea level orthometric height.

https://www.agisoft.com/downloads/geoids/

On this web page you can find geoid file.

Still struggling with the vertical accuracy in my LAS file.  I'm using NAD83/California zone 6 (usft) and NGVD29 to export the point cloud out of DJI Terra.  This gets me about 4.6  ft low on the vertical and very good on the horizontal.  Do I then transform from GRS80 to geoid 12b?

fans1c07f6ff Posted at 4-19 14:21
[view_image]

Still struggling with the vertical accuracy in my LAS file.  I'm using NAD83/California zone 6 (usft) and NGVD29 to export the point cloud out of DJI Terra.  This gets me about 4.6  ft low on the vertical and very good on the horizontal.  Do I then transform from GRS80 to geoid 12b?

Wouldn't you have added the height of the antenna from the ground. In the case of a rover it must be done, but for a base station it is necessary to put 0. Otherwise it is a guaranteed shift.