As you may know, robotics is now a trend in recycling. Robots help deal with labor problems and can be more efficient at sorting.  

However, a fast robot picker isn’t the only skill to look at. Capture efficiency, precision, adaptability, and AI are the main criteria. Let’s focus on AI.  

Every robotic company can train AI, but not anyone can read the chemistry of materials. This is where the hyperspectral takes place.  

 As seen for many years with optical sorters, hyperspectral has made its proof. Optical sorters are good, but they can only sort up to 3 fractions at the same time. Recycling robots has no limits.  

 

Contamination issues 

Using fast robots with AI will surely help sorts lot more tons per year. However, when seeing reports from MURFs, there’s still an important contamination issue. The reason is simple. Regular AI can’t make the difference between two similar packaging. 

Waste Robotics is taking care of this issue by adding hyperspectral to their scanner.  When scanning with hyperspectral, they can read the chemistry of the material and see with no doubt the difference between types of plastic. As shown on the picture, we can see on the left two identical lids. By scanning with a regular scanner, the AI won’t be able to see the difference between them both. However, with hyperspectral we can tell that the lids are made from different types of plastic. 

 

 With productors that always change their packaging, it becomes really complicated to see the difference. Both for humans and robots. 

 ’Using hyperspectral when sorting recyclables allows to reach a high and recurrent purity standard that provide more quality. When producing high-quality sorting, return on investment is much easier’’    says Eric Camirand, CEO at Waste Robotics 

 

In conclusion, it is highly recommended to look for a hyperspectral scanner when shopping for robots, so you will avoid contamination and improve the purity of your final product.