Back in April I read about this project out of Columbia University for a self-powered camera that uses the light hitting the sensor to power the camera. It's a cool idea, and could really improve the effectiveness of cameras deployed in the field. I'd be especially interested for this technology to be coupled with a mesh network so that an array of these devices could be placed throughout an area, and the images could be received and processed in near-real-time. The Buckeye Cam has wireless image transfer, but it's still too big and expensive to be a game-changer, in my opinion.
26 November 2015 3:21pm
On the subject of mesh network: what distances between nodes would you expect? I'm asking about order of magnitude: tens of meters, hundreds of meters, kilometers?
30 November 2015 11:51pm
Ideally 10s of kms within line of sight. But there's always the issue that more range means more power consumption. Optimizations like directional antennas and a low power mode that only kicks into high power when there is data to relay could help. I know ubiquiti has excellent technology in this area. Perhaps it could be leveraged here?
1 December 2015 1:21pm
I've came across ubiquiti reading about community/urban mesh networking, but (unless you mean their "non-WiFi" - meaning non-802.11 - hardware) we're talking 100s of meters between nodes there, being reasonable maximum within unlicensed band limitations (although, in the past, I personally set up a 1.3km point-to-point link using off-the-shelf 802.11b access points and simple parabolic antennas). It would be a fascinating exercise to see what is possible with commodity hardware in wildnerness. Interestingly, my past experience suggests that foliage can be equally "lethal" to radio waves as man-made structures.
As to the meshing algorithms, I recently heard a talk from a gentleman working on some of the largest "Internet of Things" pilot projects (eg. city lighting management) claiming that the (existing) self-managing meshinging protocols, although working fine in a household setting (and possibly research labs) completely failed in dozens of thousands nodes use case and that more "classic" solutions with managed infrastructure are the only way to go. I'm curious what size of the network do you have in mind? 10s of kms may be a big distance between nodes, but I guess the areas covered can be even larger?
4 July 2016 11:51am
Some camera trap manufacturers offer solar power as an add-on or option off the shelf. I agree that using the image seems like a solution looking for a problem.
Pawel Moll
Arm