[gdal-dev] Sub: Gdal, leave one out method with GCP'S (Interpolation Equation)

[Even Rouault]

Pradeep Gulla,

https://github.com/OSGeo/gdal/blob/master/autotest/gcore/transformer.py
could serve as inspiration

You could use gdal.Warp() to output to a in-memory VRT and use 
gdal.Transformer() on that

Even

> Hi All,
> 
> I am georeferencing aerial images with gcps using gdal_translate and
> gdalwarp. Before georeferencing image(warping) I would like to check
> quality and final result on a control_point cp (eg disabeld gcps).
> Therefore I would like to calculate the coordinates in destination image
> for a independent control_point based on the interpolation method (n-order,
> helmet and spline) which is NOT used by warping itself.  Afterwards compare
> positions with estimated coordinates of that cp.
> 
> Example:
> Lets say, I have 10 gcps. I take 8 gcps, using `-order n` generate a
> `model`.  Then predict destination values for rest of the two control
> points using this `model`. Finally compute shift between predicted and
> actual values for these control points.
> 
> Could you please let me know if any gdal tool does this? Can any one give
> some insight how to implement this in python?
> 
> I have checked following stack exchange issues.
> 1.
> https://gis.stackexchange.com/questions/225931/python-gdal-how-to-obtain-dx-> dy-residual-errors-for-gcps-after-gdalwarp-usage (self written n-grad
> interpolation, is there any validaton form offical gdal?)
> 2. http://osgeo-org.1560.x6.nabble.com/GCPs-to-geo-transform-td3771045.html
> 3.
> https://gis.stackexchange.com/questions/102111/preserve-gcps-after-gdalwarp


-- 
Spatialys - Geospatial professional services
http://www.spatialys.com