Este artículo está también disponible en / This post is also available in: Spanish
When we talk about artificial intelligence, we often think of it as a revolutionary and novel concept that has just arrived in our lives. However, this term was first used almost seventy years ago, in 1955 to be more precise, by John McCarthy in a workshop at Dartmouth University in the United States. Nevertheless, it is in the last decade that we have witnessed a true revolution in the advancement of intelligent technologies.
Artificial intelligence, when used under ethical standards, is poised to address many of the current challenges, including those faced by cities in Latin America and the Caribbean (LAC). Thanks to the digitization of our cities and big data, municipalities in our region are getting closer to, among many areas, reducing traffic accidents on their streets, decreasing pollution levels, or improving waste management. The latter is undoubtedly one of the main challenges for the sustainable development of the region: by detecting, controlling, and monitoring landfills in a timely manner, LAC municipalities could improve not only the environment but also the health of their residents.
In their mission to improve lives in LAC cities, the Housing and Urban Development Division of the Inter-American Development Bank (IDB) supports cities in the early detection of illegal landfills. Keep reading to learn how, thanks to artificial intelligence, the city of Mendoza in Argentina, with the support of the Bunge and Born Foundation, has taken on the issue of garbage dumps.
Open-air garbage dumps affect the development of cities
Latin America and the Caribbean produce 10% of the world’s garbage, at a rate of 1 kg per person per day. Of the total, about 35.15 million tons of this waste go uncollected daily, usually ending up in uncontrolled open-air landfills, which are often located on the outskirts of cities, near the most humble and vulnerable neighborhoods.
Waste management is a critical service, essential for urban life. Therefore, it is also a significant challenge for municipal governments worldwide, especially in developing regions like LAC. The rapid increase in waste production (it is expected to reach 671,000 tons of daily waste by 2050) in the context of limited municipal budgets makes it difficult to carry out effective waste management that preserves the environment and the health of citizens.
Open-air garbage dumps are one of the most visible aspects of the problem. These are sites where waste accumulates illegally, causing severe harm to the natural environment by degrading the quality of the soil, air, and watercourses, and underground reservoirs affected by toxic leaks. They also affect the health of those living nearby, especially the most vulnerable people. Indeed, a high incidence of respiratory diseases and heavy metal poisoning has been found among the child population in communities residing near a large landfill.
Managing these problematic areas, which are becoming increasingly common in metropolitan regions, is a massive challenge for local and national governments. Due to their informal nature, open-air garbage dumps appear and grow without the awareness of regulatory bodies. This lack of information further complicates decision-making and the creation of effective policies to improve the situation.
New technologies to solve old problems
Thanks to the development of digital technologies, it is possible to access high-resolution satellite images from anywhere in the world or even operate a small unmanned aerial vehicle (a drone) to capture aerial photographs on demand. Furthermore, new artificial intelligence techniques allow for the automated analysis of these images on a large scale, surveying the territory to detect the presence of objects of interest. So, why not use these resources to assist with artificial intelligence-based landfill monitoring?
This is the challenge that the Bunge and Born Foundation, a non-profit organization based in Argentina, took on in 2019. The goal was to develop a low-cost method for identifying and monitoring the growth of open-air landfills with high frequency using satellite or aerial images and by training an artificial intelligence-based system. From the outset, the idea was to use open-source technologies, allowing the production of a tool transferable to researchers and governments.
Large open-air landfills visible in publicly available satellite images (left) and their outline detected by an automated system (right).
The first version of the methodology was developed using the detection of large open-air landfills in the Metropolitan Region of Buenos Aires as a case study. The system, based on open technologies, proved capable of detecting 95% of the landfills reported in public records through an analysis that could be processed in a day and required nothing more than freely accessible satellite images.
In line with the goal of facilitating its reuse, the programming code that achieved these results was published and made available to the public for download and use. After the successful conclusion of the pilot, the next step was to form an alliance with a municipal government to adapt the methodology to a specific management challenge.
The Case of the City of Mendoza
In 2022, the Bunge and Born Foundation and the Environmental Subsecretary of the Municipality of Mendoza, Argentina, formed an alliance to apply this landfill detection system in the city.
In this municipality, the problem was the proliferation of small-scale garbage dumps, often referred to as “micro-dumps.” In other words, the existence of small spaces where residents and businesses in the area disposed of garbage that accumulated over time. This was causing serious ecological problems, as the rugged terrain facilitated the leaching of pollutants into watercourses downstream.
Open-air garbage dump
A specific challenge for this project was detecting landfills that, due to their smaller scale, were not visible in public satellite images. How to tackle this challenge? It was relatively simple: the methodology was modified to work with aerial photographs taken by drones instead of satellite images. The adaptation was successful, even allowing the algorithm to be trained to detect the predominant materials in the landfills (plastics, construction debris, metals, etc.) due to the higher resolution of the images.
Unmanned aerial vehicle (drone) equipped with a camera
The landfill monitoring system using artificial intelligence has allowed the city of Mendoza to complement its initial studies on waste management. The level of detail in this monitoring has facilitated the design of action plans, helping to accurately identify the most affected sites where intervention was a priority.
Left: Intervention area (about 790 ha) in the Mendoza foothills. Right: Area of intervention in the Mendoza foothills (detail)
Could your city be next to monitor landfills with artificial intelligence?
Open-air landfills and micro-dumps pose serious risks to people’s quality of life, and addressing them is a significant challenge for environmental management. The use of technologies like artificial intelligence is a step in this direction, allowing for real-time monitoring of the emergence of these landfills to take action.
The alliance between civil organizations and the public sector has been crucial to approach a more definitive solution to this problem in the case of Mendoza. In the Urban Development and Housing Division of the IDB, we are at the disposal of cities in the region to support them in following Mendoza’s example and using new technologies to monitor and control the proliferation of illegal landfills.