Illegal logging and poaching detection

Introduction

Illegal logging and poaching are major threats to the biodiverse forest and wildlife ecosystem. Timber from deep forests is harvested illegally and this accounts for a major part of deforestation activities. With this solution, we can detect illegal activities happening deep in the forests in real time.

Solution

Through this project, we aim to develop a system which can be easily placed deep inside the forest areas and can capture sounds and analyse these sounds to detect activities like using a chainsaw, motorcycles, gunshots, the sound of a falling tree etc... We use Wio Terminal to capture sounds from the surroundings and classify them using an ML model created by Edge Impulse

Model Creation in Edge Impulse

In this project, we trained and deployed an audio scene classifier with Wio Terminal and Edge Impulse. We followed this tutorial from the Seeed Studio Wiki page to create this model.

We recorded sounds directly from the laptop using the web interface provided by Edge Impulse

We used different types of audio data to classify sounds of chainsaws, gunshots, falling trees etc... We also tried experimenting with classifying forest fire sounds.

The training dataset was small and not so clean, so the model did not perform well in classifying the forest fire sound.