Computer

Aerial images and computer program locate swimming pools, water reservoirs to identify dengue-vulnerable areas

Aerial images and computer program locate swimming pools, water reservoirs to identify dengue-vulnerable areas

Brazilian researchers have developed a computer program that locates rooftop swimming pools and water tanks in aerial photographs using artificial intelligence to help identify areas vulnerable to pest infestation. Aedes aegyptithe mosquito that transmits dengue, zika, chikungunya and yellow fever.

The innovation, which can also be used as a public policy tool for the dynamic socio-economic mapping of urban areas, is the result of research and development work carried out by professionals from the University of São Paulo (USP), the Federal University of Minas Gerais (UFMG) and the Superintendence of Endemic Control of the São Paulo State Department of Health (SUCEN), within the framework of a project supported by FAPESP. An article on this subject is published in the journal PLOS ONE.

Our work initially consisted of creating a model based on aerial images and computer science to detect water reservoirs and swimming pools, and use them as a socio-economic indicator.”


Francisco Chiaravalloti Neto, last author of the article

He is a professor in the Department of Epidemiology at USP’s School of Public Health (FSP), with a first degree in engineering.

As the article notes, previous research had already shown that dengue tends to be more prevalent in inner city areas, so prevention of dengue, zika, and other mosquito-borne diseases can be made considerably more efficient using a relatively dynamic socio-economic environment. economic mapping model, especially given the long interval between population censuses in Brazil (ten years or more).

“It’s one of the first steps of a larger project,” said Chiaravalloti Neto. Among other goals, he and his team plan to detect other elements of the images and quantify the actual rates of infestation in specific areas so that they can refine and validate the model.

“We want to create a flowchart that can be used in different cities to identify risk areas without inspectors needing to call in houses, buildings and other breeding sites, as this is time consuming and a waste of human resources. taxpayer’s money.” he added.

machine learning

A previous study used artificial intelligence (AI) to detect water reservoirs and swimming pools in Belo Horizonte, capital of Minas Gerais state. The researchers first presented satellite images of the city to a computer algorithm with reservoirs and pools already identified. The deep learning program then found patterns in the images that would make detection possible anywhere, and over time it gained the ability to distinguish reservoirs and pools in the photographs on its own. .

“It’s real machine learning, a subfield of AI,” said Jefersson Alex dos Santos, a professor in UFMG’s computer science department and founder of its pattern recognition and observation lab. Earth (PATREO).

The most recent study focused on Campinas, the third largest city in the state of São Paulo in terms of population. Four areas were chosen, each with different socio-economic conditions according to the census. A drone with a high-resolution camera took aerial photographs of the areas and two datasets were created, one for water reservoirs and one for swimming pools.

The next step was to train the model and transfer the lessons learned. “We trained the model on Belo Horizonte and applied it to Campinas,” Santos said. With the images obtained in Campinas, the model became more reliable for the region, reaching accuracy rates of 90.23% and 87.53% for pools and reservoirs respectively.

Socio-economic indicator

When the algorithm was fully trained, the researchers used other images to detect reservoirs and pools in the four selected areas of Campinas and cross-referenced them with census data. The results of the analysis showed more rooftop tanks per square meter in poorer areas and more pools in richer areas.

Even these preliminary results were useful in predicting likely breeding grounds for A. aegypti. “This is not the final methodology, but it could serve as the basis for a relatively simple practical application such as the development of software to map areas of the city at high risk for dengue outbreaks,” Santos said. .

According to Chiaravalloti Neto, the model can be used for much more than controlling dengue fever and other mosquito-borne diseases. “The nation updates its socio-economic database approximately every ten years, with each population census. Our method could be used for more frequent updates, which in turn could be used to combat d ‘other diseases and problems,’ he said, adding that more markers will find their way into future studies based on aerial images, to refine the algorithms and make them even more precise.

Drone or satellite imaging?

Although aerial photographs of Campinas were taken by drone, the researchers expect the final methodology to use satellite imagery. “We used a drone because it was a pilot project, but large-scale remote sensing and scanning with drones is expensive,” said Chiaravalloti Neto.

“Also, drones have relatively low range,” Santos added. “For a large-scale project in a big city, we will need satellite images.” The Belo Horizonte survey successfully used satellite images. These must be high resolution images for the software to recognize the patterns. Access to this type of image is fortunately becoming easier, he says.

The methodology may seem expensive, but in reality, it saves time and money by avoiding the need to conduct in-person house calls to map potential breeding grounds. Instead, city public health workers can use data obtained remotely and processed by AI to select priority areas for physical inspection in a more assertive way.

Next steps

The model cannot currently detect if water tanks are properly sealed or if swimming pools are treated to prevent mosquitoes from laying eggs there. “The methodology could be refined in order to be able to distinguish between properly treated reservoirs, pools, etc., and others that may or serve as breeding grounds for the mosquito,” said Chiaravalloti Neto. Detecting such patterns and other signs of potential breeding grounds would make the algorithm even more useful to public health services.

Researchers are currently installing traps to catch mosquitoes on some 200 city blocks in Campinas. The state of the properties is carefully assessed, in particular to predict whether the mosquito is likely to breed there. Socio-economic indicators will also be analyzed. The next step will be to assess the aerial images of the areas using the logic described above to rank the risk of the presence of A. aegypti and the diseases it transmits.

“By observing these urban areas, we will build a model that prioritizes dengue control measures for the whole city, and then for the rest of Brazil,” said Chiaravalloti Neto.

In addition to FAPESP, the researchers were funded by the Serrapilheira Institute, the National Council for Scientific and Technological Development (CNPq), the Office of the Pro-Rector for Research of USP and FAPEMIG, the Agency of Minas Gerais research. SUCEN provided structural support.

Source:

Journal reference:

Cunha, HS, et al. (2021) Water Reservoir and Pool Detection Based on Remote Sensing and Deep Learning: Relationship to Socioeconomic Level and Applications in Dengue Control. PLOS ONE. doi.org/10.1371/journal.pone.0258681.