RAVI KIRAN S (kiran@cs.washington.edu)
CSE 576 PROJECT - 2 (MOSAICING)


  1. Test Sequence

    Click HERE to view full-resolution mosaic.
    Click HERE to view it in a viewer.
  2. Office sequence

    This was taken in my office with the camera mounted on the Kaidan head in a rare moment when none of my officemates were present :-). The office has windows with an outside view so there were illumination changes galore. However, the resulting output image has a uniform lighting effect although the brightness seems to have been attenuated overall. I am not entirely certain as to what causes the attenuated effect but it produces a good mosaic :-).

    Click HERE to view full-resolution mosaic.
    Click HERE to view it in a viewer.


  3. The Attack of the Aibos

    This was a handheld sequence. I had the camera very close to the ground while I took rotated taking the pictures of Aibo robot dogs in the Robotics Lab. Inspite of being a handheld sequence, the mosaic came out looking reasonably well. Unfortunately, at the last moment, the camera moved a bit and that shows up as a stitching artifact in the viewer.

    Click HERE to view full-resolution mosaic.
    Click HERE to view it in a viewer.


  4. Multiple Avatars of Ravi (Extra credit)

    This was taken in the EE1 037. I was in various positions of the scene as Pravin snapped pictures. The mosaic looks fine except for slight ghosting artifacts ( at the classroom corner and near the ladder) . Because there were identical chairs, we occassionally rotated the camera less than the desired angle, leading to the minor artifacts. This could have been avoided if the scene had some more detail and if I was centered without overlap in each frame.

    Click HERE to view full-resolution mosaic.
    Click HERE to view it in a viewer.


  5. Atrium at night - The Ferris Wheel Loop

    This fun hand-held sequence was shot by leaning over the 4th floor walkway of the atrium in Allen Center. I ignored enough vertigo to keep snapping and rotating the camera. The edges are misaligned a bit because the camera moved a bit as I maneuvered for space. As an afterthought, it would have been nicer (and stranger ) if this was shot during daylight.

    Click HERE to view full-resolution mosaic.
    Click HERE to view it in a viewer.



Other Issues
  1. The spherical warping did not work properly with the focal length and the distortion parameters given on the webpage so I used a MATLAB calibration toolkit to estimate these parameters ( available HERE ). The computed values for focal length and radial distortion parameters were f = 1070.2 k1 = -0.201 k2 = -0.43 .