Recycling: Automating the sorting and separation of e-waste.

E-waste processing will become increasingly important as a valuable source of materials for reprocessing. However, increased efficiency will only come through a technology step-change within the industry.

Stuart Watson discusses how Machine Vision and AI can be employed to improve the value that can be extracted from e-waste. Traditional bulk sorting is quick but not particularly accurate, while hand sorting is accurate but not quick. A system that employs machine learning to identify components and assign a value to the boards can work quickly and accurately with minimal human intervention. The resulting waste streams are sorted into low, medium and high value. In addition, the vision system can spot high value components and separate these for recovery. We believe that the combination of machine vision and machine learning has the potential to create additional value for e-waste recyclers.


At Cambridge Consultants we’ve been thinking about printed circuit boards and what happens to these and all the products that they’ve been built into get throw away and sent for recycling. These boards can have a huge variation in value largely down to all the different materials that they can be made from, from gold, palladium through to the plain copper that we more commonly associate with them.

There’s a number of different methods for sorting these boards into different values, but the two most common ones are simply sorting them based upon the product they come out from – recyclers like this because it cheap, but it’s not particularly accurate, down to sorting each board individually. Now this requires an individual who has experience of the field to look at a board, assess it, and then put it in an appropriate bin. It’s accurate, but it’s not particularly quick and requires someone who has years of experience to make a judgement call.

We wanted to combine these two methods to make a single system that was both fast and accurate. We did this through using the two digital technologies of machine vision and machine learning to process each board individually and then apply algorithms that have been trained to value each board, it can do this at the great speeds required for a modern recycling plant.

Another advantage is that it can identify individual components. So if a specific capacitor type, or a specific processor chip has a high reuse second-hand value in the industry, that can then be segregated from the main waste stream to add additional value and profit to the recycling company.

We think that these are good examples of how digital technology can add additional efficiencies and value to waste handling in the recycling industry.


Stuart Watson
Principal Engineer

Stuart is a chartered mechanical engineer within the Analytical Engineering Group. Prior to joining Cambridge Consultants, he spent several years in the aerospace and consumer industries with hands on experience of precision mechanisms and mass production.