Ultrapurple
Hyperspectral imaging rig test
One of my problems with attempting hyperspectral imaging has always been making sure that the alignment between the different images is at least reasonable. Here I have mounted a Therm-App thermal imager on the hot shoe of a Nikon D3100 DSLR. This gives a predictable and repeatable relationship between the two images, although there are still parallax errors (particularly for close items).
In this staged shot the thermal image on the phone screen was taken a few moments earlier and the rig repositioned so that the live view on the camera screen looked right. The parallax error at this close range was very significant.
This setup opens the door to putting several bodies together (a pair of IS PROs for UV and NIR, the D3100 for visible and the Therm-App for LWIR), giving coverage in one image from 14µm up to ~350nm, albeit with a MWIR gap from ~7-1.2µm. (I'll keep my eyes open on eBay for an ad MWIR camera...)
Of the two glasses of water, the one on the left was straight from the cold tap and the one in the right had had a dash of warm water added so that it was a little over lukewarm. The difference is clear in the thermal image but of course indistinguishable in the visible image.
The 6.8mm Therm-App lens was found to have a field of view approximately equivalent to a 28mm lens on a 35mm full frame. It does however exhibit quite strong optical distortions, which will have to be compensated for before thermal images can be overlaid on a visible or other image. (This assumes that the visible camera's lens is distortion-free – a big assumption...).
Comments are warmly welcomed.
For more thermal images covering a diverse range of subjects please visit (and join!) the Therm-App (and others) thermal imaging group at www.flickr.com/groups/therm-app-users/
Hyperspectral imaging rig test
One of my problems with attempting hyperspectral imaging has always been making sure that the alignment between the different images is at least reasonable. Here I have mounted a Therm-App thermal imager on the hot shoe of a Nikon D3100 DSLR. This gives a predictable and repeatable relationship between the two images, although there are still parallax errors (particularly for close items).
In this staged shot the thermal image on the phone screen was taken a few moments earlier and the rig repositioned so that the live view on the camera screen looked right. The parallax error at this close range was very significant.
This setup opens the door to putting several bodies together (a pair of IS PROs for UV and NIR, the D3100 for visible and the Therm-App for LWIR), giving coverage in one image from 14µm up to ~350nm, albeit with a MWIR gap from ~7-1.2µm. (I'll keep my eyes open on eBay for an ad MWIR camera...)
Of the two glasses of water, the one on the left was straight from the cold tap and the one in the right had had a dash of warm water added so that it was a little over lukewarm. The difference is clear in the thermal image but of course indistinguishable in the visible image.
The 6.8mm Therm-App lens was found to have a field of view approximately equivalent to a 28mm lens on a 35mm full frame. It does however exhibit quite strong optical distortions, which will have to be compensated for before thermal images can be overlaid on a visible or other image. (This assumes that the visible camera's lens is distortion-free – a big assumption...).
Comments are warmly welcomed.
For more thermal images covering a diverse range of subjects please visit (and join!) the Therm-App (and others) thermal imaging group at www.flickr.com/groups/therm-app-users/