--penalty-weight for convert_pinhole_model

[jchollingsworth]

Hi... I have a feature request ("urghh" I hear you say)... but I hope this is the simplest of tweaks:

Would it be possible to add the option --penalty-weight (from cam2rpc) to convert_pinhole_model, when converting to an RPC distortion camera.

I find this option very useful in cam2rpc, but convert_pinhole_model seems too sensitive to high frequency variations at times.

Cheers, J

[jchollingsworth]

A small additional point...
Given the distortion varies more for a given number of pixels at the edges of an opticalbar camera, a uniform --sample-spacing might not capture well the variations at the edges of the image(?). Perhaps this explains why pinholeRPC doesn't well approximate opticalbar at the edges (as seen using mapprojected images).

Would it be possible to densify the sampling at the the left/right edges of the camera, using less points across the middle of the camera?

Cheers, J

(P.S. I know you guys won't have time to look at this, but I put the points out there for when you do, and also in the hope these are very simple things to adjust in the source code(?).)

[oleg-alexandrov]

James, thanks. I don't know when we will get to this, but these are sensible suggestions. Glad you are pushing the tools to the limit. We hope to at some point make ASP very easy to compile and maybe features that you dearly want you could add yourself. :) Hopefully later this year.