Advice computer vision for Recycling Machine

Advice for the computer vision software development of a recycling machine

Waste processing is quite a challenge in the consumer sector. Although municipal authorities are doing a lot to encourage waste separation and thus possible reuse of raw materials, plastic in particular is a major problem. Plastic as waste only acquires value as a raw material when it succeeds in producing so-called mono-flows. A Dutch startup with an elaborated concept for a Recycling Machine for these mono streams approached us for advice regarding the machine software and computer vision. Because of secrecy we can only describe in very global terms what we have done,

What is the problem that needs to be solved?

.

Client has chosen the second option to start with the option to be able to apply the 3rd at a later stage in an industrial version of a consumer machine.

Advice

We have issued a broad and concrete advice, in which the following elements are relatively unique:

• The use of a Computer Vision camera to be able to read barcodes as e.g. with Google Lens

.

• The use of a Jetson Nano platform allows you to develop both fully embedded operating software as well as smart connected through public APIs

.

• The future use of public APIs (i.o.) for plastic recognition based on form and label identification (also driven by CV)

.

• The use of an existing Dutch initiative for crowdsourced obtaining a valuable database barcode -> plastic type packaging
• The use of Software Control & Distribution to ensure that enrichment of an embedded database can easily be implemented across multiple machines, making the concept scalable

.

Mindmap

We start this kind of advice or prototyping trajectories, e.g. with a mind map on which issues and possible solution directions are shown. We did the same for this assignment.

Translated with www.DeepL.com/Translator (free version)