Madrid will employ AI to spy on waste deposited outside of containers

The municipal cleaning company says that this will speed up the resolution of waste incidents

FCC Medio Ambiente, the municipal street cleaning company of Madrid, announced that it will start testing the use of an AI system to detect improperly placed trash outside of waste bins. The aim of the system will be to speed up the removal of such waste by vehicles of the company in order to ensure better street hygiene.

The experimental phase will run in three neighbourhoods of southern Madrid: Carabanchel, Usera and Villaverde making them pioneering spaces in all of Spain when it comes to the use of this “spying” technology in the field of urban maintenance.

AI will assist the waste inspectors in their job

The system works by having AI sensors attached to the vehicles of the company’s inspectors circulating on the streets. Previously, these people had to manually pay attention to the bulky pieces of trash deposited outside of the containers, such as furniture and boxes, and note them down for reporting.

The AI system, which is equipped with machine learning algorithms, has been progressively trained to recognize different objects, both large and small, and immediately communicate their presence and location to a centralized monitoring hub. From there, the system will find where the nearest waste removal vehicle is located and assign it the task of picking up the bulk.

FCC claims that this will speed up the whole resolution process by 50%. Previously, the waste patrolling rounds would take five hours, but now they can be completed in only two hours. And the best thing is that the driving inspectors can concentrate on other tasks since the AI system will essentially do the more mundane work.

With its effective implementation, the AI model will continue to be perfected since each recording will improve its initial design and detection capacity through real-time training. Furthermore, to comply with data protection legislation, a program has been incorporated that automatically pixelates the people and vehicles that the cameras capture during the inspection journey.