Opened 13 years ago

Closed 13 years ago

#2650 closed defect (worksforme)

3rd order GCP fitting

Reported by: rjc1 Owned by: warmerdam
Priority: highest Milestone: 1.5.4
Component: default Version: 1.5.3
Severity: normal Keywords: 3rd order fitting
Cc:

Description

The documentation says 3rd order GCP fitting is supported. But it produces distorted results. After looking through the GDALCreateGCPTransformer code: if nRequiredOrder is set to 0, the code sets it to 2nd order even though there are 10 or more GCPs. There is also a comment in the source code saying 3rd order is unstable. I downloaded GDAL for a NIST project expecting it to work upto 3rd order. While we are at it 5th order is highly desirable.

Change History (1)

comment:1 Changed 13 years ago by warmerdam

Resolution: worksforme
Severity: criticalnormal
Status: newclosed

I believe it is still possible to create 3rd order polynomial transformers by explicitly passing in 3 as the order. If zero is passed in (automatic) nothing higher than 2 will be used since higher orders are pretty unstable.

If you find it is not possible to pass the order in as 3 (to GDALCreateGenImgProjTransformer()) then please reopen this ticket, indicating details.

Note: See TracTickets for help on using tickets.