Re: [hugin-ptx] Re: Masking inside hugin
wow really nice feature..thanks for adding.. On Tue, Feb 16, 2010 at 7:44 PM, Bart van Andel bavanan...@gmail.comwrote: On 16 feb, 09:35, Bruno Postle br...@postle.net wrote: http://www.flickr.com/photos/36383...@n00/4359863335/ This will save a lot of work: instead of first outputting all the remapped images, editing masks by hand using a photo editor like GIMP or Photoshop, and then blending, everything can now be done from inside Hugin. Very nice work indeed! -- Bart -- 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.comhugin-ptx%2bunsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/hugin-ptx -- _ Asst Manager IT Wazir Ali Industries Limited Lahore -- 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
Re: [hugin-ptx] Re: Masking inside hugin
Seems to work fine, thanks for this :-) Tim http://img94.imageshack.us/img94/7807/posmasktest.jpg http://img130.imageshack.us/img130/2101/negmasktest.jpg On 16 February 2010 08:37, sebastien delcoigne sebastien.delcoi...@gmail.com wrote: Thanks Thomas, It's a very practical feature to have directly inside Hugin. I can't wait to try it. -- Sebastien On Tue, Feb 16, 2010 at 9:22 AM, Carl von Einem c...@einem.net wrote: Standard (negative) masking: Make sure that certain areas of a source image (a partial human body stepping into the frame, or parts of the panohead) don't make it into the stitched image while you have enough better background in another frame (e.g. a handheld nadir shot). The 'crop' tab is comparable but only allows to mask the outside of a rectangle or a circle. Positive masking: comparing two overlapping frames one feature might be in both but looks better in frame B. You want to make sure that enblend uses that nicer part so you can apply positive masking on that wanted part. Bruno's tutorial shows both techniques using seperate vector masks: http://hugin.sourceforge.net/tutorials/enblend-svg/en.shtml Jan Martin schrieb am 16.02.10 07:58: This might be obvious to you, I ask nevertheless: What is masking good for? -- 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.comhugin-ptx%2bunsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/hugin-ptx -- Sébastien -- 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.comhugin-ptx%2bunsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/hugin-ptx -- 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
Re: [hugin-ptx] Image overlap.
Thanks, Bruno, for a useful and informative post. I didn't know that! I suppose some allowance also needs to be made for seam wander. However, it does suggest the point that higher resolution sensors may well need less overlap, the seam has absolute dimensions. It didn't occur to me that points closer to the image centre are necessary for calculating lens distortion. It might also be the case that correcting distortion towards the middle of the image is less critical than getting it right in the overlapping regions. Bruno rightly points out that working from a lens profile is the way to go, I completely agree. A minor problem can occur with zoom lenses with profiles; focal length may well vary a little between series but, I suspect this wouldn't cause too much trouble in practice. Emad makes the valid point that wider lenses suffer greater distortion and require greater overlap. Using a custom profile with a wide lens still allows for lower overlap, limited in practice by the cropping needed. I would suggest that lens profiles are definitely the way to go. Then overlap is dictated by the needs of the blender and output cropping. In general then, wider lenses need greater overlap, principally to avoid excess cropping. Longer lens need less overlap, constrained more by the blending seam. Still, nothing is achieved by overdoing the overlap; 30% is still too much! John -- 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
[hugin-ptx] Re: What is hugin's best fit?
Some notes from my experience. I got a Panoraus last year and was very disappointed with the initial results. I don't have the numbers for the CP errors to hand by they were quite high. After I while I realised the problem was that I had attached the Panoraus to my normal camera tripod's head. This had a lot of slack when the camera was tilted up/down. With luck I found I could unscrew the tripod's head and attach the Panoraus directly to the base of tripod. This was much better and I now normally find the worst-case cp error is 2. It sounds to me that you are making the calibration process too hard. My calibration technique was to use two nails, one near the lense (30cm) the other further away (2m). I them moved the camera back on the Panoraus slide, and rotated the camera to see the two nails move relative to each other. Then moved the camera back and repeat. I just visually assessed the movement on the cameras view-finder. Repeating these steps until the movement was minimised. My other comment is that with my lense the FOV number from the EXIF data is too high (96degree) when I allow Hugin to optimise this it is normally reduced to 94degrees.. This significantly reduces the average CP error. Maybe these comments will help you. Regards Stephen On Feb 16, 7:35 am, icysubdweller rosyrobo...@inwa.net wrote: I've seen hugin report good fit and bad fit after optimization. Is there anything better? Does it ever say, Excellent fit? I bought myself a pano head for Christmas, and today I had a chance to play with it and try to calibrate it. I shot a sequence of 2-row x 4- pic panos of the side of my house, moving the camera out the upper rail 3mm at a time. I'm allowing plenty of overlap between shots, easily upwards of 30% of image width/height on all seams. I saw the errors reported by hugin start out large, get smaller and smaller, then start getting big again as I slid the camera out the rail. So then I shot another sequence of panos around the setting with the smallest error, this time moving the camera 1mm per sequence. The best result I got was Mean error = 3.3 and max error = 17.1. There were still some visible problem spots in the image, but I figured, Eh, automatic CP generation, I could do better by hand. To verify, I brought my setup inside and shot an indoor pano using the best settings I had discovered outdoors. The initial results were very good, on par with what I got outside: mean error = 2.8 and max error = 13.8 using automatic CP generation. But still some visible problems. So now I've spent the last 3 hours tweaking the control points. The image dislocations seem to move around, but not predictably, and are never completely gone. Qualitatively, 4 hours of work playing with positions of CPs, making a few tentative forays into optimizing lens params, etc., hasn't changed the result at all. If I go hunt around the image and count the problems, the number and size of the dislocations are always approximately the same, no matter what I do. Does anyone have any suggestions where to go with this? What kind of error levels are needed to produce a perfect result (which I define as not being able to spot any stitching errors in the final pano)? Is this a calibration issue of the pano head? Play in the tripod/head/ pano head setup... would it cause this? If my error rates seem low enough, what could be other causes of my issues? Complex lens distortions not modeled/correctable by hugin? Or does it just take more practice than this, and this is all easily explained by user inexperience (in which case, what are the beginner problems I'm likely overlooking)? Also, along a different line: I'm shooting with an Olympus E-620 and a Zuiko ED 12-60mm lens at 12mm, which is 24mm in 35mm-equivalent terms (2x crop factor). It's a medium-wide lens, and image elements in the corners of the image can be rather rotated from one image to the next. Often I find the largest errors reported for CPs are in the corners of the images. Autopano-sift-c tends to always find CPs towards the center of the images; usually it's my hand-placed CPs that are out in the corners, and which report large errors. I often can't fine-tune the points because the fine-tuner often moves one of the two points somewhere else in the image where it finds a better match. Yet the points are visually dead-on when I place them, on well-defined image features. ??? Does that give anyone a clue what I might be experiencing? Thanks for any ideas... -- 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
Re: [hugin-ptx] [OSX] A Hugin 2010.1.0 svn4984 bundle with Thomas Modes Mask Editor and Pablo's patent free CPdetector included
On Wed, Feb 17, 2010 at 12:39 PM, Harry van der Wolf hvdw...@gmail.com wrote: Please find attached to this mail a screenshot of a general purpose configuration for patfree-panomatic. Note that you should not add a path for the program name as it is an internal tool. I'm not seeing the patfree-panomatic appearing in PreferencesControl Point Detectors. I tried, against your suggestion, to add a path to the program, but all I got was a wxExecute error as I recall when it tried to run. Should this new detector be showing up when Hugin launches? iMac, 20 inch, G5 under OS 10.5.4. Steve -- 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
Re: [hugin-ptx] Re: What is hugin's best fit?
Good comment, I think the point to emphasise is that parallax correction can be fairly approximate, and still provide good results. Parallax accomodation is only critical when producing images of near fields, such as interiors at high resolution, and with longer focal lengths. Still, do everything you can to eliminate Px, and create lens profiles for your chosen focal lengths. I agree that, in most cases, a simple hands-on approach to parallax correction, will work well. An example might inform: I determined a parallax correction for my 10mm and EOS camera of 100mm, by trial and error. I've gained very good results, I built my lens profile around this correction (eliminate parallax before creating a lens profile). I learnt subsequently, that I was about 7mm out, but couldn't detect any image problems. Hugin handled it all. Take lots of pictures. Improve by increments. Every defect is a gem. John -- 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
Re: [hugin-ptx] Image overlap.
On Wed 17-Feb-2010 at 13:10 +0500, Emad ud din Butt wrote: very nice informative data.But Is it due to distortion factor in wide angle lenses that required more overlapping. Because when you shoot with medium to tele lenses there is no distortion and stitch is easily done. what do you say about this? You need different overlap depending on whether you are calibrating lens parameters or stitching. Often in Hugin you are doing both at the same time so it isn't easy to give good general advice. -- Bruno -- 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
[hugin-ptx] Re: Masking inside hugin
Awesome work! Very useful to the workflow. Could you make it use the roll parameter? I used some portrait photos, but in the mask(and crop) tab they were in landscape, while in CP editor they were properly in portrait orientation It would also be great if some CP generators made use of masking. PTmender would also rock if it supported masking. Thanks you the feature! -- 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
Re: [hugin-ptx] [OSX] A Hugin 2010.1.0 svn4984 bundle with Thomas Modes Mask Editor and Pablo's patent free CPdetector included
You have to set it upin the Preferences window: - 'Control Point Detectors' tab - New... - Description: patfree-panomatic (or something that makes sense for you) - Type: All images at once - In the first ('Detector') tab: - choose One step detector from the drop down list - type patfree-panomatic (without the quotes) in the 'Program' field - use these arguments: --sieve1size 100 --sieve2size 2 -o %o %i Cheers, Carl phartz...@gmail.com schrieb am 17.02.10 23:27: On Wed, Feb 17, 2010 at 12:39 PM, Harry van der Wolf hvdw...@gmail.com wrote: Please find attached to this mail a screenshot of a general purpose configuration for patfree-panomatic. Note that you should not add a path for the program name as it is an internal tool. I'm not seeing the patfree-panomatic appearing in PreferencesControl Point Detectors. I tried, against your suggestion, to add a path to the program, but all I got was a wxExecute error as I recall when it tried to run. Should this new detector be showing up when Hugin launches? iMac, 20 inch, G5 under OS 10.5.4. Steve -- 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