When material composition and market demand evolves, so does our AI. Our waste recognition library just grew from 89 to 111 categories.
The latest taxonomy update helps facility operators target the most valuable plastics they extract in their waste streams. It also improves our AI’s performance across the board: composition data becomes more specific with each update, which translates to more accurate weight estimates.
Greyparrot Analyzer’s new recognition categories are tailored to the material that our customers — and the reprocessors they sell to — want most. Together, they’ll help facilities maximise product purity, and meet the growing demand for high quality recycled plastics.
The expansion unlocks capabilities that enables facilities to charge a premium for purer plastic bales:
Separate high-value plastics by opacity and colour
Users can now track PET and HDPE objects based on opacity and colour — two variables that have a major impact on their value.
Our customers are using the expansion to:
💰 Command premium prices for PET and clear containers with dedicated recognition classes covering transparent, translucent and opaque plastics.
🎨 Create a higher-value stream of HDPE with new distinctions between white and coloured HDPE objects.
Reprocessors pay a premium for transparent PET and white or natural HDPE recyclates, which are more versatile as secondary resources. The ability to separate based on those qualities gives sorting facilities a competitive edge in a market that is looking for recycled material with the quality to contend with virgin plastics.
Track clear containers in more detail, and focus in on food-grade
Seven new classes for clear pots, tubs and trays offer even more insight into some of the most important commodities in the waste stream. Operators can now determine whether objects are:
- Lids
- Tubs
- Trays
- Pouches
- Netting
…or even fragments of larger objects.
More context also means clearer differentiation between food-grade and non-food grade material, another key driver of bale price.
Target common contaminants
We consulted with our customers around the world to identify the contaminants posing a threat to their most valuable product lines.
The resulting classes enable them to adapt as non-target material appears in real-time, and plan ahead to ensure fewer contaminants enter their facilities in the first place:
💿 CDs and blister packs regularly appear on PET lines, so we’ve introduced dedicated classes to flag their presence.
🧪 Tubes of all kinds — from crisp cans to toothpaste, cosmetics and silicone sealants — are now on our radar.
🧤 Latex gloves and shoes now have dedicated recognition classes that help minimise contamination on flexible plastic and textile lines.
Sorting is just as much about what you remove as what you keep in. By spotlighting objects that continue to appear in otherwise-pure bales, we’re helping facility operators remove another hurdle to high-quality recycled plastics.
Responding to the growing demand for recycled plastic
Between extended producer responsibility (EPR) fee modulation and Europe’s PPWR scheme, demand for resources rPET and rHDPE is set to boom.
By expanding our taxonomy, we’re empowering facility operators to target and recover high value plastics at a critical moment in the circular transition.
Our industry-leading waste taxonomy now extends to over 111 types of waste, making it one of the most detailed AI waste analytics systems on the market. Explore the full recognition library here 👇
