Inspiration
The rampant destruction of forests due to illegal logging is one of the most pressing environmental challenges of our time. This activity not only depletes natural resources but also fuels organized crime and human exploitation. Our team was inspired to develop a solution that leverages the power of artificial intelligence, neural networks, and cloud technologies to actively combat illegal logging and its associated human rights violations. We envisioned SIGHT AI as a tool that could give governments, NGOs, and environmental activists real-time insight into where and how illegal logging is happening, and help expose related activities such as forced labor.
What it does
IGHT AI uses neural networks to classify satellite imagery and detect illegal logging activities. Our system scans and analyzes vast regions of forest using AWS cloud storage to handle large amounts of data. Through LangChain, we incorporate news articles and local reports, connecting suspected illegal logging sites with potential forced labor incidents in the region. Users can access an interactive map, built with React, that overlays areas of concern, providing authorities with actionable data in real time.
How we built it
Neural Network for Classification: We developed a neural network model trained on synthesized satellite data to classify regions of forest as either legally or illegally logged. Flask API: We built the backend using Flask, providing endpoints that process satellite images and return predictions about illegal logging activities. Cloud Storage on AWS: We used AWS to store and process satellite imagery at scale, ensuring smooth access to large datasets. React Frontend: Our frontend, built with React, allows users to explore a global map with AI-based detection overlays, enabling real-time monitoring of illegal logging activity. LangChain Integration: LangChain helps us analyze and pull relevant news articles from the surrounding areas of detected illegal logging, linking these sites to possible human exploitation, such as forced labor.
Challenges we ran into
Data Processing: Handling and processing large amounts of satellite imagery data was a significant challenge, requiring optimization of our neural network and storage solutions. Model Training: Training a neural network to accurately classify illegal logging sites was a complex task, involving extensive experimentation with datasets and fine-tuning. LangChain Integration: Ensuring accurate linkage between satellite-detected logging activities and related news articles through LangChain presented challenges in filtering relevant information from local news sources.
Accomplishments that we're proud of
Successfully developing an AI model capable of detecting illegal logging through satellite imagery classification. Building a seamless system that integrates Flask, React, and AWS to store, process, and present data in real-time. Effectively linking environmental monitoring with news analysis using LangChain, allowing us to uncover possible connections to human exploitation.
What we learned
The complexity of dealing with satellite imagery data and optimizing neural networks for large-scale classification. The importance of cloud architecture in handling massive datasets, particularly in environmental monitoring applications. How to integrate NLP tools like LangChain to bring contextual insights into an environmental project, revealing the interconnectedness between illegal logging and forced labor.
What's next for Sight AI
Expanding coverage to include more regions and enhancing our AI model’s accuracy by training on diverse datasets. Developing prediction capabilities to forecast potential illegal logging hotspots based on historical data trends. Improving the integration with LangChain to include more sources and better identify human rights violations. Partnering with government bodies and NGOs to deploy SIGHT AI for real-world use, enabling them to take action against illegal logging and its broader implications.
Log in or sign up for Devpost to join the conversation.