For years, recycling packaging depended mainly on manual processes and people. But what if artificial intelligence could help decide how to recycle better? This is the question that frames the fourth episode of the series marking the 30th anniversary of Sociedade Ponto Verde.
After recalling the major milestones, the shift from quantity to quality of recycled materials and the social impact of packaging recycling, “Packaging and Innovation Always on Point” reaches the moment when technology takes center stage in the process. Today, artificial intelligence is already transforming the way packaging waste is identified, separated and valued.
A camera that sees beyond the ordinary
One of the examples comes from Sentinel, a Portuguese company specialized in inspection solutions and intelligent vision for industry, which realized that systems of this kind were missing in packaging recycling.
Thus was born Sentinel Smart Sorting, an innovative solution that combines computer vision and artificial intelligence. With a camera that sees beyond the ordinary, this system automatically identifies contaminants in the waste sorting and guides the robots to remove them, enabling efficient packaging recycling. Supported by Sociedade Ponto Verde and in partnership with Resinorte, Valorização e Tratamento de Resíduos Sólidos S.A, this technology enables a more efficient and accessible circular economy for all.
The big challenge is variability: the plastics that pass through in a week can be very different from those of the following week. The team trained its AI models to maintain good efficiency day after day. The most interesting result is economic: by using more conventional color cameras, the technology achieves results similar to much more expensive systems, with a lower investment. In practice, this means more automation on the lines, more waste treated, and better recovery rates.
Algorithms to Put Recycling Points in the Right Place
Artificial intelligence does not stop at the sorting line. As part of RE.SOURCE, the open innovation program run by Sociedade Ponto Verde in partnership with the Beta-i consultancy, Algar faced a very concrete challenge: how to accelerate placing ecopoints where they are needed, in areas with high population but few collection points available.
With the Portuguese startup NILG.AI, a tool was developed that cross-references open data with the organization’s internal data (collection histories, population density, infrastructure) to identify the best locations. The results speak for themselves: the time to place an ecopoint was reduced by more than 80%, from around two weeks to a few hours, and the acceptance rate of the suggestions by municipalities also rose. A project that was awarded the “Data Changemaker of the Year” prize.
From a system that reacts to a system that anticipates
The same logic is reaching selective collection. Instead of collecting by the calendar, routes are now adjusted to the real behavior of people. As the episode sums up, “before collection was calendar, today it is increasingly behavior,” and that is possible only because there is data.
With projects like the one Algar has already completed in RE.SOURCE, it is possible to forecast fill levels, automatically optimize daily routes, and adjust circuits in near real time. The system stops reacting to what happens and begins to anticipate what will happen: a system that observes, learns, anticipates and evolves, gradually turning an operational infrastructure into a true system of continuous knowledge about the materials.
Tecnology and sustainability hand in hand
Manual and automatic processes, people and machines: it is from this combination that the future of packaging recycling is built. As the episode concludes, separating better and recovering more is the objective, and when intelligence enters the system, packaging recycling also evolves.
