Compostable packaging requires compatible collection and industrial treatment. Reliable identification can help processors separate unsuitable plastics before they reduce the quality of finished compost.
Why identification matters
Some compostable plastics are intended to break down under controlled industrial composting conditions and may be collected with food waste where an appropriate system exists. Conventional plastics do not behave in the same way. If they enter an industrial composting stream, fragments may remain in the finished material.
The research examined whether hyperspectral imaging could support faster classification of plastics entering this process. The technique records how a material reflects light across many narrow wavelength bands, producing information that can help distinguish polymers that may look similar to the eye.
Materials and method
The experimental work used short-wave infrared hyperspectral imaging alongside several data pre-processing techniques. A chemometric classification approach—Partial Least Squares Discriminant Analysis—was then used to interpret the spectral information.
The tested material groups included compostable and conventional polymers:
- PLA
- PBAT
- PHA
- PP
- PET
- LDPE
Samples included clean materials, soil-affected pieces and plastics recovered during industrial composting. This allowed the researchers to examine performance beyond ideal laboratory conditions.
Reported findings
reported classification accuracy for the specified virgin polymers and soil-affected pieces larger than 20 mm by 20 mm.
reported accuracy when distinguishing the tested clean large microplastics from soil-affected samples.
reported identification accuracy for tested plastics and large microplastics found within compost.
The tests also showed that real processing conditions matter. Identification performance varied with sample darkness, size, colour, thickness and the degree of contamination. These factors limit how directly laboratory results can be transferred to a working facility.
Context: A compostable label does not mean an item will break down in every environment. Performance depends on the material specification, certification conditions, collection route and treatment facility.
What the work suggests
Imaging-based identification could become one part of a broader quality-control system for industrial composting. It may help distinguish selected polymers and locate larger fragments, but the reported variations indicate that equipment, operating thresholds and validation with representative waste streams remain important.
The wider system still requires clear product labelling, appropriate collection instructions and coordination between packaging design and local treatment capacity. Identification technology is a useful intervention, not a substitute for those connected decisions.