From Roadways to Recovery: repurposing ADAS Vision for Global Recycling

Advanced Driver Assistance Systems (ADAS) are finding an unlikely home: the local recycling center. The same sensor fusion and computer vision technologies used to detect pedestrians and lane markings are now being repurposed to identify used clothing, plastics, and metals on high-speed conveyor belts. This "Industrial ADAS" is enabling recycling facilities to operate at speeds and accuracy levels that human sorters simply cannot match.

By applying deep learning models trained on millions of images of waste, these systems can distinguish between different grades of plastic or types of fabric in milliseconds. When the AI identifies a high-value recyclable, it triggers robotic actuators to pluck the item from the stream. This reduces contamination in the recycling chain and turns what was once "trash" into a high-purity raw material for manufacturing.

This application demonstrates the versatility of perception software developed for the automotive industry. As the cost of high-resolution cameras and edge-AI processing continues to drop, we can expect to see ADAS-like "safety and detection" layers integrated into all forms of heavy industry, creating a cleaner, more efficient circular economy.

Source: NVIDIA Blog