The Parallax Pano Problem

I am waiting to finish off the bottom of my mounting box to test out the setup and see if I can make 3/8th of a panorama. I am aware of the problems with parallax, which is when objects are viewed at different angles. This is an advantage to us humans as it allows us to determine the relative closeness of things because we have two eyes. With the pano-cam I’m building though, it will potentially be a big problem due to the artifacts created in the stitching process.

Panoguide says this on the subject of panoramas and parallax:

If everything in the scene is approximately the same distance from the lens, or very far away, the parallax effect will be elminated or at least minimized. This is why hand-held panoramas from a mountain top tend to work quite well – everything is far away from the camera, and because the distances are very large compared to the tiny distance between the nodal point and the point of rotation, parallax is negligible.

So, as people get closer the the cameras, the more parallax will be apparent in the resulting panoramas. This is unfortunately the trade off of taking higher resolution photos with off the shelf equipment. Instead of 10-12MP panoramas with a 360 lens, I expect to be getting 40+ megapixel panoramas, which will give more of the experience of being there by being able to zoom in on parts of the photo. I absolutely love the gigapan project for this reason, and I’m hoping users will appreciate the higher resolution too.

So that being said, I won’t really know what it will look like until I run tests, which should be very soon. My hope is that it will be “good enough”, because it’s going to be awhile until I’ll be able to afford going the other one-camera-with-a-360-lense route.