🌳 An AI-powered solution to help save the environment
A Non-profit trying to help save our Planet
From monitoring forest fires and illegal logging to saving farmlands. Tree Planet is driven to help save the environment with a sophisticated technology driven approach.
Why Bother?
The Problem
Our Environment is very fragile, with issues like forest fires and logging contributing to its destruction. Whilst developed countries have the resources to try and combat the issue, many low-income countries aren't able to implement such systems. Furthermore, farming and crop yields are at risk since theft is a critical problem for independent farmers in these nations.
Built on A Robust Network
Stay Updated all the time
Our Network Architecture uses a mix of LoRaWan® and other communication methods so that you never lose connection of an Instance (device). To learn more about our ESP-powered instances and our Network Architecture please click below.
Keep track of everything all in one place
Our easy-to-use dashboard helps you locate, study and monitor any disturbances that might need your intervention. Options to flag, close, and open cases are readily available as well as the ability to request a new audio sample. Any data flagged by the AI that is true or false will be collected for training purposes to further improve the model's reliability.
AI with 96% Accuracy
Powered by a 16-layer CNN architecture
Google Maps Integration
Easily locate and interact with instances on our map view
How it works
We offer a secure and reliable SaaS product to organisations or private entities who would benefit from our Forest Fire and/or Illegal Logging Detection Software.
Donations pay for the operational costs
Your Donations help more farmers or disenfranchised communities (in lower income countries) to implement our solution.
Protects the environment and crops
More organisations use our donation-backed solution to help monitor and save trees from fires and logging, and farmland from theft.
Reduces Pollution and Increases Yield rates
Preventing fires and deforestation reduces harmful pollution increasing the standard of living, whilst protecting farmland from theft increases crop yields and contributes positively to the economy.
Up to 96% detection accuracy
Developed on a dataset of around 40GB, our 16 layer Convolutional Neural Network can easily identify audio based on MFCC and Spectogram representations during extraction. The analysis of the data had been dockerised and deployed on Azure and Google's Vertex Ai Platform. The data is then served by an API onto the frontend.
Devices that cost as low as $5 save the world
Our prototypes started with a very simple architecture with ESP-32 boards that communicate via Wi-Fi. However in much more complex use cases this isn't sufficient, so technologies like the LoRaWan protocol would be ideal despite the added component costs. This technology eliminates the need of a WiFi connection, which is useful in forests. Though this is not enough since there isn't peer to peer communication between the instances, so we have developed an extra layer of communication using the onboard RF sensor of these much more complex, and expensive devices.
Our devices run on Solar Power, and store the excess power overnight in a LiPo battery.
Easy-to-use dashboard packed with features
Our dashboard offers many features, allowing the user to add/delete instances, set up custom user roles, view instances on a map, get notified if audio is flagged by the AI, and much more.