Support Website Contact Support Blog

vertical shift between point cloud and DSM data

I’m confirming the accuracy of using UAS and Pix4DMapper for our road projects in anticipation of using aerial photography for our hard surface mapping.  I was able to successfully create a point cloud and the necessary files to import a DSM into Civil 3D.  At first I wasn’t too pleased with my results.  I used 21 GCP’s over a 1.5 mile project and had them located with RTK GPS.  I selected the proper state plane coordinates for my output files however; when I imported the “” over a project which was completed with total station, I found that my elevation was a consistent 0.40’ off.  But, if I reviewed the coordinate in the rayCLoud editor they are right on!  If I applied a manual -0.40’ to the dsm in civil3D everything matched great!  I was surprised to see my data within a 0.07’ of our total station locations.  This is an amazing result, but I’m concerned about why I had to apply a manual shift.

My question is:  How come in the rayCLoud editor my elevations are perfect, but yet when I import the dsm surface into CIvil3D those same elevations are shifted 0.40’?   

UPDATE:  after further review the vertical error was found to be in an unadjusted GPS pole by the surveyor.  All of his GPS locations on the GCP’s were 0.24’ higher than actual.  This brings the results into range depicted on the error report in Pix4D.

Hello Tony,

Thank you for the update.

Good to know that you got the expected results.

I think you can find useful this blog post. It compares the accuracy obtained with GCPs and RTK/PPK geotags.

In a few days, there will be a new one analyzing the accuracy of the point cloud depending on the GCP distribution which could be of your interest too.


Hi again,

This is the blog post I was talking about in my previous post:

Do more GCPs equal more accurate drone maps?