3 ms·

Neat. I wonder how much work it will take before it can hover. I just spent too much time on YouTube watching people crash drones (often into people). I coul

Neat. I wonder how much work it will take before it can hover. I just spent too much time on YouTube watching people crash drones (often into people). I could see some applications for this filming at crowded venues once it's more developed.

andrewflnr 10y ago
Putting a camera on any flapping-wing platform is going to require a heck of a lot of image stabilization, to say the least. You would probably have to put the camera on a neck like a heron's.
- jayjay71 10y ago
  I thought about that shortly after posting my comment. I could still see it being used by consumers for low quality videos, or maybe for medium quality depending on how much can be salvaged from software stabilization. But then just thinking about the weight alone for decent cameras and how much more they weigh than this entire platform! I see it mostly as a cool research project right now. Like most research projects it will likely never see mass commercialization, but it's still really cool. And maybe something they learn from this will be useful for something else entirely. I also actually know one of the professors who worked on this, as he taught my first programming course way back when. He just loves to build cool robots.
  - andrewflnr 10y ago
    Oh yeah, I'm not saying the project is anything short of awesome!
- Turing_Machine 10y ago
  A long neck would be be one way. There are others. Eagles don't have have particularly long necks (nor do bats, for that matter), but they nonetheless make their livings by homing in on rapidly-moving targets that are performing active evasion. Clearly the eagles and bats have some way of compensating for the wing motion. I know there's been a lot of work on how bats do internal image processing (sonar images, in their case). Not sure about eagles.
  - andrewflnr 10y ago
    Animals don't need to form a stable image the same way a digital camera does. They (and we) are already compiling a mental visual model from a lot of pretty sparse samples (you may have heard that we only have fully detailed vision in a very small portion of our visual field). I guess filtering out motion is pretty easy to add on to that. I think existing digital image stabilization software does something similar.

andrewflnr 10y ago

Putting a camera on any flapping-wing platform is going to require a heck of a lot of image stabilization, to say the least. You would probably have to put the camera on a neck like a heron's.

jayjay71 10y ago
I thought about that shortly after posting my comment. I could still see it being used by consumers for low quality videos, or maybe for medium quality depending on how much can be salvaged from software stabilization. But then just thinking about the weight alone for decent cameras and how much more they weigh than this entire platform! I see it mostly as a cool research project right now. Like most research projects it will likely never see mass commercialization, but it's still really cool. And maybe something they learn from this will be useful for something else entirely. I also actually know one of the professors who worked on this, as he taught my first programming course way back when. He just loves to build cool robots.
- andrewflnr 10y ago
  Oh yeah, I'm not saying the project is anything short of awesome!
Turing_Machine 10y ago
A long neck would be be one way. There are others. Eagles don't have have particularly long necks (nor do bats, for that matter), but they nonetheless make their livings by homing in on rapidly-moving targets that are performing active evasion. Clearly the eagles and bats have some way of compensating for the wing motion. I know there's been a lot of work on how bats do internal image processing (sonar images, in their case). Not sure about eagles.
- andrewflnr 10y ago
  Animals don't need to form a stable image the same way a digital camera does. They (and we) are already compiling a mental visual model from a lot of pretty sparse samples (you may have heard that we only have fully detailed vision in a very small portion of our visual field). I guess filtering out motion is pretty easy to add on to that. I think existing digital image stabilization software does something similar.