Image
stability and registration should be achieved primarily
by fabricating a stable platform. Inclusion of visible
registration marks within the frame would allow for re-alignment
of images. Some off-the-shelf network camera systems may
be able to be adapted, but in any event, the technology for
all of these aspects is readily available, or easily developed.
Large format CMOS chips may be preferable to CCDs, which for
standard consumer and video applications require the ablity
to capture at short exposures, which we do not require. Also,
future image standards will be much higher resolution than is
currently available and should be anticipated as much as possible.
A telemetry system with sensors to report the status of capture
unit to the network should also be included.
Consistancy
of lighting effects from frame to frame will be maintained over
time by the use of sun angle and spot metering of key image
elements. An algorithym will select moments for optimal exposures,
from which the frames most consistant with the previous days'
images will be selected for upload to the network from the memory
cache .
Keeping
the lens clean over long periods may be a challenge, but could
be accomplished in a number of ways. The most maintenance free
might be using an electric motor to periodically spin a disk
of glass serving as the first element, and thereby disperse
any contaminants. This would not maintain a perfectly clean
surface, but should be able to maintain a consistent view.
A
perfectly clean surface could be achieved with an roll of moving
plastic film that would advance periodically when triggered
by a photo sensitive sensor detecting the scattering caused
by contaminants on the surface. An air or water jet could also
be used, but these would all require some differing amounts
of periodic maintenance.