On Fri, Sep 02, 2016 at 04:06:34PM -0700, Abrimaal wrote: > Let's talk here what are the optimal settings (projection, CP detector > etc.) for panoramas of untypical objects or photo techniques, to minimize > the number of attempts. > > 1. Analog B/W photos. Two lucky shots taken from the same position, one > after one. Both images were printed on paper and after 40 years scanned. > Photos were scanned in 8-bit greyscale, 600 dpi. The paper was the same > size but not in a perfect condition (slightly folded). I was trying all CP > detectors, each of them failed. I removed all the CP, added them manually > and again the same error.
If you added CPs manually, don't try to run CPfind again. Just move on to the optimization stage. > 2. Scenes with people or other moving objects. How many people were > crippled or beheaded stitching in Hugin. Some other software has a > possibility of graphical adjustment of the stitch lines. Is it possible in > Hugin? Sometimes I take a photo for a panorama with a passing car or a > human. I want to make two versions of the panorama - with and without the > object. I agree with Gnome, use the masking tools. I generally only use exclude regions to cut out obviously-wrong areas, as that leaves the blender open to more options. > 3. Panoramas of cars and other reflective surfaces. The reason "why" is > simple, often cars stand too close to each other or to a wall, to take a > full photo. > The CP detectors find points in reflections of other cars, of buildings in > the car surface, not in the "hard" parts and high contrasts. After deleting > all the false points and matching "hard" points, there is always at least > one error in the panorama. The camera was rotated around the nodal point as > it was possible from the hand. How to minimize the risk of wrong stitch? Manually inspect the CPs in highly-reflective areas. Make sure you re-run the optimization before stitching, and pay attention to CPs that have a high (over 1.0) error. If this error will not go away even with full distortion parameter optimization and all good CPs, then there is parallax. There is no real fix for parallax besides re-shooting the photos or getting creative with photo manipulation. > 4. Partially solved. Straightening the car side views. Ideally a car should > be photographed from the maximal distance, the smallest field of view, what > means loss of quality. To straighten the side view I make copy of the > photo, then I load both images and use Align Stack detector. I add the > V-lines manually on buildings, not on > the car. The rectilinear projection usually makes the image stretched at > the edges, what is visible that the wheels are not round. The adjustable > Panini general seems to > be the best choice, but sometimes the final image requires re-scaling in > another editor, it is too tall compared to its length. I'm afraid I don't really understand what you're doing in this example. Can you provide an example picture set / pano of this? --Sean -- A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ --- You received this message because you are subscribed to the Google Groups "hugin and other free panoramic software" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/hugin-ptx/20160903151737.GA8655%40wheatley.localdomain. For more options, visit https://groups.google.com/d/optout.
