IdentifAI Core (Brann forebygging)

Offentlig prosjektsammendrag

Stellai develops an AI-based detection system for hazardous objects in waste streams, such as batteries and gas canisters, which pose a high fire risk in recycling facilities. The goal of the project is to create a robust, real-time system that can operate without pre-trained datasets and adapt to constantly changing waste environments. The main challenge is the complexity and unpredictability of visual waste data, which makes traditional AI methods ineffective. To address this, the project applies a systematic R&D approach combining real-world data collection, continuous testing in operational settings, and development of adaptive computer vision models using methods such as anomaly detection and self-supervised learning. The result will be a new product that helps reduce fire risk and improve safety in an industry with limited access to modern detection tools.