h a l f b a k e r yCrust or bust.
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
| |
Do you think such a system could distinguish between carnivorous sharks, their more placid cousins and dolphins? |
|
| |
//various noxious chemicals released into the water// They do this in the Irish Sea and there isn't a shark in sight. |
|
| |
Different sizes. If we're talking about great whites, tigers, bulls and the like I think they're all fairly large, larger than a harmless Caribbean reef shark or a dolphin, and much smaller than a basking or whale shark. Speeds are most likely different too. It could be engineered to tell the difference I'm sure. |
|
| |
The speed and movement of a large shark would be quite different to a diver or even a school of fish. The system could be tested in known shark feeding waters, maybe training it up using a kind of nervous net system to allow it to develop its own intelligence. |
|
| |
I thought this was going to be security sharks like security blankets. Well, Calvin had a security *tiger*. |
|
| |
IIRC, nervous nets can't train (unless you consider the original construction the training), they don't have any memory, you can't create nonlinear decision boundaries, and they scale horribly. i.e. they're not really good for anything except Tilden's PR. Instead use a multilayer perceptron, radial basis function or a support vector machine. Maybe a hidden markov model or a Boltzmann Zipper if you want to incorporate time series info like swimming patterns. Not as sexy as nervous nets, but at least it'll work better than chance. |
|
| |
prometheus, if you read 'neural net' for 'nervous net' I think you'll get Nadir's intentiion. Unless something has happened in the years since I studied them, a multilayer perceptron is basically the same as a neural net. I agree Tilden's nets aren't really applicable. |
|
| |
I read something some years ago about soldiers who went swimming in the ocean off of South Africa. Every so often, they'd throw a couple grenades far out into the surf to scare the sharks away. I don't know if it really worked or not, but they never mentioned anyone getting bitten. |
|
| |
That could easily have the opposite effect to the one intended, [Guncrazy], as the grenades are likely to blow up other marine life, the smell of which would attract sharks. |
|
| |
Gee, I was planning a system very similar to this one. So it's definitely a + :-). |
|
| |
Something that could perhaps help in distinguishing sharks from dolphins is the fact that sharks have no bones, only cartilage. I guess sonar reflection would be somewhat different. |
|
| |
As a deterrent I was told that in those islands where they have "dive with sharks" underwater excursions they sometimes hit sharks in the nose when they get more frisky (I guess these are not realy big sharks). So, for bigger sharks we could have a minisub equiped with a shootable boxing glove that would hit sharks in the nose until they went away. When the shark turned away from the beach the minisub would shout "and don't come back!!!"... hmm, maybe I should post this as an idea. Can I? |
|
| |
Another way to distinguish sharks from other sea life is based on the shape. Dangerous sharks usually have very pointy noses and the fins in the chest area have very sharp angles in relation to a dolphin's, for instance. Same applies for dorsal fins. |
|
| |
// they'd throw a couple grenades far out into the surf to scare the sharks away // |
|
| |
I once read a story of a diver who used to scream to scare sharks away. He was surprised when tried it on a different country and didn't work. The reason was preciselly that in that area they used dinamite for the same effect. |
|
| |