Skip to main content

Waste analytics: artificial intelligence at the service of waste recognition

Automation for tracking, and the analysis of waste and its correct sorting

Advances in artificial intelligence continue unabated with increasingly diversified areas of application.
With a view to pushing constantly towards technological innovation, we have created an innovation project aimed at testing the use of hardware and software based on state-of-the-art technologies for the monitoring of waste as it passes through our plastic sorting facilities.


Whitin the A2A Group's sorting facilities, both those specifically for plastics and those dedicated to paper sorting, it is necessary to carry out frequent product analysis aimed at evaluating both new recovery opportunities and optimizing the performance of the plant itself.
At present, material analysis is carried out manually by the plant operators with spot checks at variable intervals and over a period of time. This can lead to an ongoing lack of understanding of what is happening inside the facility itself, resulting in the possible loss in value of some of the material that is not effectively intercepted by the sorting technologies available in the facility.


Through the experimentation of technology proposed by the Greyparrot start-up, we have successfully tested an automatic monitoring system of waste flows inside our plastic sorting facilities (CSS). Specifically, the system is made up of a set of cameras, a computer vision model based on artificial intelligence and a platform designed to manage and share the data acquired during monitoring. The goal is to deliver real-time analytics of materials such as different types of plastic as they pass along the conveyor belts in our facilities in order to increase our level of knowledge of the operation of the plant and enhance its functioning.

Currently, this technology is being used on a PET sorting line in the Muggiano facility.


This technology can allow optimization of plant performance with resulting increases in the value of the material in transit.

In detail, the main benefits concern these areas:

  • Automated and continuous monitoring of waste flows moving along the plant sorting lines, thus reducing the need for manual analysis;
  • Increase in the level of knowledge of the materials going through the facility and greater visibility with respect to the efficiency of the sorting processes. This, when fully operational, can lead to higher earnings thanks to the ability to intercept and make economically advantageous use of greater quantities of material;
  • Reduction of the risk associated with the potential sanctions of the COREPLA consortium for non-compliance with the purity requirements of waste material flows (eg. PET, polypropylene, etc.);
  • Possible integration with other systems, both software or hardware and machinery, already used in the facility or which could be used at a later date by means of APIs, providing cognitive AI (automated sorting machines, for example);
  • Correlation between changes in composition and specific factors related to the time of year or sorting staff differences.


The experimentation has shown that the system is able to provide results with great precision, limiting error in the recognition of the different types of materials in transit to less than 3%. The system also proved to be able to monitor the composition of flows on different timescales (from month to minutes), highlighting the importance of continuous monitoring and providing feedback at various levels.

Innovation projects for the Circular Economy