Spaced Out :: 12 2011

Capturing aerial imagery from KAP can be a hit and miss affair (although far less so with digital cameras) and therefore setting the camera up carefully prior to imaging is essential. For oblique images there are fewer problems because there is a lot of light coming in to the camera and you can utilise whatever shutter speed/aperture you want. That’s not to say its not without problems, but that automatic camera settings can often get great shots.

When you’re trying to obtain straight verticals from KAP in order to generate DEMs (see my paper) then things become more difficult. It reminds me of a more general quote from my undergraduate days (thanks Lindsay):

"In aerial survey we have to take a large number of photographs to a set plan, from a moving platform, at a great height and at a low temperature. Alone or collectively these conditions are foreign to most photography. Despite this the photogrammetrist requires excellent definition of a low contrast object, further reduced by poor atmospheric conditions."

Which about sums things up….. OK, its not quite as severe from a kite, but the requirements for high definition photos of a low contrast object remain the same. This is even more so when we consider that we want to collect stereo imagery: the photos need to be good. So when it comes to the camera how do we actually set it up?? My paper linked above has the following paragraph in it:

"The photogrammetric use of aerial imagery requires sharp definition and this is controlled on the camera by focus, shutter speed, aperture and ISO settings. With flying heights in excess of 50m, a focus setting of infinity is used. The focus ring is normally taped and auto-focus setting switched off, this eradicates changes to the focal length during the flight. As the kite is usually moving during image capture, fast shutter speeds are needed. Field tests suggest speeds in excess of 1/500 s are required. In addition an aperture of at least F8 (for a wide angle DSLR) is preferable, to allow sufficient light to enter the camera system whilst minimizing lens distortions. In most field situations an ISO setting of at least 400 is necessary for normal UK daylight conditions. With automatic digital cameras, these stipulations mean that the ISO is set prior to image capture with the camera in 'aperture priority' mode. Prior testing may be required in order to ensure shutter speeds are fast enough."

So there we have it: tape the focal ring at infinity, put the camera in aperture priority and maintain an aperture of at least F8. The camera will auto-set the shutter speed but we actually need speeds in excess of 1/500s. This can be problematic as the only light entering the camera is reflected off the Earth’s surface and for low reflectance features (e.g. vegetation) this can be quite low. One thing I didn’t explore in the original paper was the AutoISO feature that is available on the D70 and pretty much every Nikon DSLR since. Yes, it automatically sets the ISO for you and there is a nice layman’s article here. Nikon Support also offer some sage advice. In short though, AutoISO increases the ISO speed (aka sensitivity) of the sensor in order to expose the photo correctly. The brilliant aspect when shooting in aperture priority is that you can set the threshold shutter speed at which it should kick in. So in our case we could specify 1/500s to always make sure photos are sharp.

This is no panacea though… yes it effectively adds 3 or so f-stops but you increase the noise in the image, particularly once you get over 1000. And whilst we want sharp photos for photogrammetry, noise won’t help things. Clearly some more experimentation needed here, but AutoISO is well worth enabling.

We also mustn’t forget the ambient atmospheric conditions at the time of image capture. We want sun… yes, but more than that we want texture in the image. If we are interested in enhancing topography then shadows are helpful. A low angled sun can significantly improve texture but again you lose sunlight which affects exposure. Its a careful balance which you rarely get right all the time!

P.S. And don’t forget to turn the auto-shut off to the longest possible time. You don’t want the camera to power down after 30 s!!

Pleiades 1 has now been launched and give Europe similar commercial imaging capabilities as the US. Slightly more detail (but older) here showing the 50cm coverage over four bands. Also confirms the future launch of SPOT 6 and 7 to provide a high resolution constellation family. The BBC also provided good coverage as this was somewhat of an unusual launch as it came from a Russian Soyuz rocket launched from French Guiana. See the launch in all its glory!

NASAs Earth Observatory have posted the first global image from VIIRS on NPP. This is a great start to demonstrate the capabilities of VIIRS; and for those doing introductory courses in remote sensing the accompanying notes nicely explain why we have sun-synchronous orbits and why there is variation in the imaging of the global composite.

Just as the Journal of Maps is transitioning over to a subscription based model we see the government increasingly becoming interested in open access publishing. The lead in The Guardian is:

“The government has signalled a revolution in scientific publishing by throwing its weight behind the idea that all publicly funded scientific research must be published in open-access journals.”

which is verging on the side of sensationalist, but does reflect a genuine interest by government to become more transparent and make outputs of funded research more widely available. I am a big supporter of open-access as it allows all to access research work, but as The Guardian notes, journals add significant value through peer-review and the actual publication. Someone as to do this and it has to be funded. They note the physics research community moving funding from subscription to author-pays (something we used at JoM), but consensus varies by discipline. The actual cost of publishing a journal article is quite high (somewhere over $1000) and for many disciplines with very low amounts of research funding this is not sustainable. And it begs the question about where to publish non-funded research (again, varying by discipline, and often in very significant quantities).

So if the author doesn’t pay and the library doesn’t pay, who does? Well that’s a good question and why the subscription model remains the most popular as it is the closest solution. Many open access journals are directly funded by learned societies who also have subscription based journals, whilst learned societies also rely upon journal subscriptions to fund research activities. I don’t know what the mix between commercial and learned societies is, but it would be interesting to investigate. I don’t think anyone would argue with reducing excessive profits by some publishers, but there has to be a publishing solution in place that is fit-for-purpose.

Some of you may know that I’m a big fan of Carbonite, the online backup facility. Yes it charges, but for unlimited storage and a real no-hassle backup it is a bargain at $59. Which in my book is pretty good for the 85Gb I’ve got stored there. Whilst it’s really billed as online backup, it is a “webdrive” and the web interface to access files is pretty reasonable. Of course, there is also now an Android app to do this which is simple, small and fully functional. Very handy to be able to grab a file whilst out and about.