[bug] Panorama Pipeline - Multi-Camera Rig

[jeffreyianwilson]

Attempting to stitch fisheye images from multi-camera rig, PanoramaPrepareImages throws [error] Only one intrinsic allowed (4 found)

Input is 4 raw DNGs from NavvisVLX lidar scanner

[fabiencastan]

Could you provide one sample dataset?

[andrewjafa]

images-part1.zip images-part2.zip

Hi Fabien,

I'm having the exact same issue as above (same lidar scanner too) - I have attached four DNG images (in 2 zips due to file size limits)

Cheers

[andrewjafa]

I have been able to work around this by deleting 3 of the intrinsic entries then copying the remaining intrinsic ID to each of the viewpoints

An alternate way was to use the "Multi Cam Rig" method, which makes sense because the Navvis VLX is a multi-camera rig. This succeeds with PanoramaPrepareImages, however then fails for me at on the following node (PanoramaInit, 'Error parsing input file')

[andrewjafa]

I tried again with a different set of images with the 'rig' method and PanoramaInit succeeded - however it now fails at PanoramaEstimation, citing 'Rigs are not currently supported in PanoramaEstimation' - is there a way to work around this? Apologies I know this wasn't the original question (and wasn't mine either) but I'm struggling to find any information

[Erol444]

Hi @andrewjafa , did you have any success with multi-cam rig & panorama images pipeline? I'm going through the similar "battle" as yourself, and found your comments useful:)

[andrewjafa]

Unfortunately not, I couldn't find a workaround for PanoramaEstimation