Battery Sorting

Challenge

Sorting and removing batteries from mixed waste on a conveyor belt is a critical but difficult task. Batteries pose both fire and environmental hazards if they remain in the waste stream, but manual sorting is highly labor-intensive, repetitive, and prone to error. Human operators must stay alert for long periods to identify small batteries of different shapes, colors, and sizes mixed among a large volume of waste, increasing the risk of missed items. The process also involves safety concerns, as handling batteries incorrectly can cause leaks or sparks.

Solution

Our solution deploys an AI-powered vision system integrated with a robot at IVAR's waste management facility in Norway. Using a high-resolution industrial camera combined with advanced AI detection models, the system identifies batteries among other waste items in real time. Once a battery is detected, the AI calculates its position on the conveyor and sends precise pick coordinates to the robot. The robot then safely picks and removes the battery from the waste stream, ensuring it is separated for proper recycling.

This automated approach improves both safety and efficiency by reducing the need for manual sorting, minimizing the risk of fires, and ensuring a higher accuracy of battery removal. The system is designed to handle variable conveyor speeds and adapt to the unpredictable shapes and appearances of waste, making it robust for industrial-scale operation.

Results

• Detection accuracy above 95% — batteries of varying shapes, sizes, and colors reliably identified among mixed waste streams
• Replaced 2 full-time manual sorting roles on the line, eliminating repetitive and hazardous handling of waste
• Fire incidents linked to battery contamination dropped to zero within the first months of operation