forest_spy

Front end
Data analysis time domain
spectrogram

Inspiration

I was always enthusiastic on creating a better place for all beings in the world. As humans we are part of nature and take every measure to protect and conserve the Earth.

What it does

In this project we are trying to develop a system that alerts the national park management to be aware of any illegal activities happening in the parks and conservatory areas.

How we built it

We are using a neural network and signal processing techniques to identify and analyze different sounds in deep forests. We used "Environmental Sound Classification 50" dataset in our Convolutional Neural Network model to classify different sounds.

Challenges we ran into

It was challenging to integrate backend with front end.

Accomplishments that we're proud of

We have successfully trained the neural network completely and developed the back end needed. We also designed a nice front end for our system.

What we learned

This is our first full stack development and we learnt about lot of new things on the way. We learnt about the data processing using Tensorflow, frequency domain signal analysis using scientific mathematical tools like matlab.