71–80 of 128 posts
Someone more familiar with satellites can correct me, but from my understanding you should be able to detect deforestation from real-time satellite data. You alert on changes in the level of greenery in the images. The indicator might trail behind detecting chainsaw noises, but it wouldn't require the mass distribution of detector devices.
Excited to see how these systems are integrated into prevention / intervention and other applications for these systems beyond detecting illegal logging.
Have the creators of this project considered the privacy implications of capturing, recording, or analyzing audio from the rainforest? What happens if people are walking near a microphone and a conversation is picked up by the microphones? I realize that the audio is probably analyzed for the sound of chainsaws, and then thrown away, but there's still the potential that such a system could be misused. So just wonderi…
Have the creators of this project considered the privacy implications of capturing, recording, or analyzing audio from the rainforest? What happens if people are walking near a microphone and a conversation is picked up by the microphones? I realize that the audio is probably analyzed for the sound of chainsaws, and then thrown away, but there's still the potential that such a system could be misused. So just wonderi…
EDIT: Seems that another commenter is aware of a company doing something like this: http://www.shotspotter.com/company
As an experiment - my neighborhood in India has a problem with loud motorbikes that are illegally modified. Hypothetically if I wanted to use this approach (there are probably simpler ones in this specific case), where do I start?
Do I just start recording gigabytes of audio data every day and start doing some kind of FFT on the data and manually tagging cases of loud bikes?
After that I suspect I can use one of those ML cheat sheets to figure out what is the best model for this kind of thing and then I just send it data via the TensorFlow library?
On the surface of it it feels like I'd be spending hours of processing time converting the audio data and hours of manual tagging before I get anywhere close to a solution. Is there a quicker or smarter way to go about this?