will this (calibrate lens gui / detecting lines) work on source (rect / fish) images only? or also a final pano in equirect projection?
On Wednesday, May 18, 2011 9:03:59 PM UTC+2, T. Modes wrote: > > > > On 18 Mai, 04:18, Tduell <tdu...@iinet.net.au> wrote: > > Hullo All, > > I have just committed a patch to the default branch which installs > > calibrate_lens_gui.desktop on Linux. It works OK here on my Fedora > > build. > > Any Linux users who want to try/test the lens calibration tool might > > want to grab the tip. > > Some hints for calibrate_lens_gui: It is based on lens_calibrate (Tims > GSoC project from 2008) > You need images with straight lines. > Load them, check the focal length, crop factor and lens type. Then try > to detect lines. If lines were found the should be painted as green > lines in the preview (Black lines are invalid lines.) If no lines were > found try to change the parameters: In my test the most critical > parameters is minimum line length: this is the minimum length a line > must have to be detected as line. Is is given in the fraction to the > image width (So try with smaller values 0.1 or 0.05) > The edge detection scale and thresold influence how the edges are > detected. You can switch the preview to Edge detection to see the > direct influence (need to press find lines again). > If enough lines are found you can optimize the lens parameters. After > it switch the preview to corrected and check is the result is correct. > If so save the lens parameters to a file. > > Thomas > > PS: What should be happen with the original calibrate_lens? > I'm for removing it: First calibrate_lens is using it's own lens model > which is not compatible with libpanos model. So the output does not > help any user to correct an image (In several mails I asked for a > explanation of the parmeters but nobody could explain it.) Second the > source code is ugly: mixing global variable with local variables. So > it is complicate to reuse the code or fix bugs. The main parts I > refactored into an own lib (huginlines) from where they can now easily > used. Also it pulls in several dependcies which are not used in the > current state (They are only used when activated some deactivated > code). Removing calibrate_lens would also allow to remove ANN from src/ > foreign. > What's your opinion? Has somebody used calibrate_lens and vote for > keep it? -- You received this message because you are subscribed to the Google Groups "Hugin and other free panoramic software" group. A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ To post to this group, send email to hugin-ptx@googlegroups.com To unsubscribe from this group, send email to hugin-ptx+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/hugin-ptx
