Waste sorting and AI glossary

The key terms of sorting, waste stream characterization and computer vision, explained simply.

Baling press: Machine that compacts sorted materials into bales for storage and transport. A dangerous object crushed in a press (battery, gas cylinder) can cause a fire or an explosion. See fire prevention.
Computer vision: Field of artificial intelligence that enables a machine to analyze images: recognizing an object, its material or its position. In a sorting center, it identifies waste on the conveyors in real time.
Conveyor: Belt that carries waste from one sorting step to the next. Wasoria sensors are mounted on gantries above the conveyors to analyze every object that passes.
Dangerous object: Waste that creates a risk of fire, explosion or machine damage when it ends up on a sorting line: lithium batteries, gas cylinders, nitrous oxide cartridges. These objects cause nearly €100 million in damage every year in France to sorting centers and incinerators.
Incompressible object: Rigid object that resists compression, such as a gas cylinder or a nitrous oxide cartridge. It damages sorting equipment and presses, and can explode if it still contains gas.
Lithium-ion battery: Rechargeable battery found in phones, e-cigarettes, cordless tools and scooters. Crushed or punctured during sorting, it can heat up suddenly and start a fire: it is a frequent source of fires in sorting centers. See dangerous object detection.
Machine learning: Artificial intelligence method in which a model learns to recognize objects from a large number of examples, instead of following hand-written rules. Wasoria trains its models on its own database of waste images, built in a sorting center.
Nitrous oxide (N2O): Gas sold in cartridges for culinary siphons, often misused as a recreational product. Empty cartridges and cylinders end up in recycling bins: incompressible, they damage machines and can explode in presses or furnaces. See their automated extraction.
Purity: Share of the target material in a sorted flow, expressed as a percentage. Mass purity is calculated on weight rather than on the number of objects. It determines the resale price of sorted materials.
Selective collection: Separate collection of recyclable waste, such as packaging and paper. Sorting errors by residents bring dangerous objects into it, such as gas cylinders or batteries.
Sorting center: Industrial facility that receives waste from selective collection (packaging, paper) and separates it by material, using machines and sorting operators, before sending it to recycling channels.
Sorting residue: Waste set aside during sorting because it does not belong to the target recycling channel. It is then sent to energy recovery or disposal.
Thermal camera: Sensor that measures the temperature of objects from a distance and turns it into an image. Mounted above a conveyor, it spots a hot spot or a fire outbreak before it spreads. The Wasoria Box combines a thermal camera with a high-resolution color camera.
Waste characterization: Analysis of the quality and composition of an incoming or outgoing waste flow: the share of each material category is measured in a sample. It is used to monitor collection quality, sorting performance and the compliance of sorted materials. See Wasoria Characterization.