Detecting Deforestation from Satellite Images
Abstract
We are currently facing a climate change crisis, with the world producing 52 billion tonnes of CO2 every year. One of the contributors to climate change is deforestation. Trees are natural carbon sinks that capture C02 from the atmosphere. When cutting down trees, we are reducing the amount of C02 that nature is capturing from the atmosphere and puts more of the burden on us. Beside reducing carbon capturing trees, deforestation also disrupts the ecosystem and can could potentially have negative effects on wild life, food sources, water reserves, and more. One of the ways to combat deforestation is early detection. Satellite images can help us survelliance forest from the sky, but survillence camera on it's own is not usingful unless we can automatic the detection progress using machine learing. Using a dataset of satellite images of the Amazon Rainforest, a convoluntional neural network can be trained to detect deforestation from satellite imagery that can aid in the early and automatic detection of deforestation to allow relevant authorities to take action promptly to prevent deforestation and protect our natural carbon sinks.