Map Processing and calibration with Survey

I am having some difficulty aligning our maps. Our company currently surveys Targets in with high accuracy RTK gps and we then fly the mission and pickup the targets. Our issue lies with the EPSG codes. We do not survey in any particular coordinate system. Local ground coordinates only. Even if we were to use an EPSG code coordinate system some of our jobs are not close enough to published control points to get a baseline for the particular coordinate system. All we need to do is overlay our maps to scale over our surveys in CAD. When you scale the image with 2 control points as expected the map is distorted and the other control points do not fit. The only thing i have been able to do is use the rubbersheet command in Civil3d to manipulate the image based on control point coordinates versus selected points on the images. This works extremely well. 2d maps only of course. but all of our survey points match up with the map image within centimeters. i need to eliminate this Rubbersheeting process in Cad. so my question is…is there a way to calibrate the map in both directions so it is scaled properly to all of the control points. Using only 2d local coordinates? I need this to be done during the processing stage in drone deploy so when i output the image to our drafting department everything overlays perfectly. Some of our survey crews do not even use GPS so it is not possible to get into an “epsg” code system. we need the distances on the processed map between control points to match the measured horizontal distances on the ground…thats it!


1 Like

This is something that just about everyone I have spoken with is struggling with and I would suggest a holistic review of your cad workflow and how the drone data is going to be integrated tightly with it.

The correct way to do it is through drawing setup with the correct projection and scale factor information and the TFW file that comes with the orthomosaic will line it up. There shouldn’t be any special functions or plug-ins involved. The two things that are the most common causes are the scale factor from survey, the drawing has been shifted or never started on the correct coordinate system in the first place.

If you would like to share a representative cad file and orthomosaic dataset I would be happy to take a look at it. Just PM me a download link if it is sensitive.

Do you know you’re correct epsg?

Hi Sean

If you post-process (GPS PPK), take a look at EZSurv software, it allow to use local grid projection.
If you know at least 3 points in Lat/Lon/H and their equivalent values on the grid X, Y, Z well you can create a local grid for further post-processing.
Once your local grid is created, GCPs or image position can be computed using your local grid projection. Therefore all output can be made to grid (X, Y H on the grid).

so how are people processing maps when no survey control is available? rural areas for example there are no 3TM/UTM survey reference markers for miles.

It depends on what your goal is. If it is to get absolute local accuracy then they are not.

You can get relative accuracy with a PPK setup which can later be calibrated (adjusted) to fit an elevation benchmark if needed, but the horizontal still may not match the land surveyed ground unless the files are in a true State Plane coordinate system.

Depending the post-processing software used, some like EZSurv, allows to create local grid (to overlay with your CAD point positions).

To create a local grid you need to identify at least 3 existing points and give their equivalence Lat/Lon/H and grid coord X/Y/Z.

Once the gird is created, you can post-process your GPS data and output the results to grid coordinates.

If you have a GPS on the drone, you can also post-process the drone trajectory, then interpolate each photo position (based on photo timestamp), the outputted photo position can also be outputted to the grid to perfectly overlay with your existing data.

While I really like the pre-PPK GCP idea, it is not the same as letting the photogrammetry processing use the GCP’s to localize the map. I think it is great that you can emulate the process I am doing in 3 steps, but users should know the difference.