Mapping by aerial images of exposed water tanks and swimming pools can help control dengue

A computer program capable of identifying, from aerial images, water tanks on roofs or slabs and swimming pools in open areas was developed by Brazilian researchers with the aid of artificial intelligence tools. The proposal is to use this type of image as an indicator of areas especially vulnerable to mosquito infestations. Aedes aegypti, which transmits diseases such as dengue, zika and chikungunya. In addition, the strategy emerges as a potential alternative for a dynamic socioeconomic mapping of cities — a gain for different public policies.

The research, supported by Fapesp, was conducted by professionals from USP (University of São Paulo), UFMG (Federal University of Minas Gerais) and Sucen (Superintendence for Endemic Disease Control) of the São Paulo State Health Department. The results were published in the journal PLOS ONE.

“What we did at this first moment was to create a model that is based on aerial images and computer science to detect water tanks and swimming pools, and use them as a socioeconomic indicator”, says engineer Francisco Chiaravalloti Neto, professor from the Department of Epidemiology, Faculty of Public Health, USP.

In the published article, he and his colleagues point out that previous surveys already showed that poor areas of municipalities are often more predisposed to dengue. That is, using a relatively dynamic socioeconomic status update model —especially compared to the Census, which is done every ten years and subject to delays—would help prioritize dengue, zika, and chikungunya prevention efforts.

“This is one of the first steps of a broader project”, highlights Chiaravalloti Neto. Among other things, the group aims to incorporate other elements to be detected in the images and to quantify the real infestation rates of the Aedes aegypti in a given region to refine and validate the model. “We hope to create a flowchart that can be applied in different cities to find areas at risk without the need for home visits, a practice that takes a lot of time and public money”, says Chiaravalloti Neto.

machine learning

In a previous study, the group had already used artificial intelligence to identify water tanks and swimming pools in Belo Horizonte (MG). The researchers started by presenting these satellite images of the mining town to a computer algorithm and pointing out which ones had these facilities. Through a process of deep learning (or deep learning), the program started to identify patterns in the images that indicated the presence of a pool or water tank. Over time, the system was able to differentiate these structures in the photos on its own.

“It’s really a machine learning process, a subarea of ​​artificial intelligence”, explains Jefersson Alex dos Santos, professor at the Department of Computer Science at UFMG and founder of the Laboratory for Recognition and Patterns of Earth Observation.

For the current research, the professionals delimited four regions of Campinas characterized by different socioeconomic contexts, according to the IBGE (Brazilian Institute of Geography and Statistics). A drone with a high-resolution camera flew over these areas and took a series of pictures. So, a database was created for water tanks and another for swimming pools.

The next step was to perform the transfer of learning technique. “We trained that model in Belo Horizonte and applied it in Campinas”, explains Santos. With the images obtained in the city of São Paulo, the models became more reliable for the region, reaching an accuracy of 90.23% for the detection of swimming pools and 87.53% for the detection of exposed water tanks.

Socioeconomic indicator

With the algorithm properly trained, the researchers used other images to calculate the concentration of exposed water tanks and pools in those four previously selected regions of Campinas. When crossing this information with IBGE data, it was noticed that the socioeconomic indexes were lower in areas with a greater concentration of water tanks and higher where there were more swimming pools.

As less structured regions are more prone to infestation of the Aedes aegypti, this model would already help in the fight against the diseases propagated by it. “Although it is not yet the final methodology, we can already think of a practical and relatively simple use of developing software for large-scale use, with the objective of mapping neighborhoods with greater risk of dengue outbreak”, emphasizes Santos.

Chiaravalloti Neto points out that the models created could be useful beyond the control of dengue, zika and chikungunya: “Updating the socioeconomic indices of different parts of Brazil takes place every ten years, with the Census. in a more agile way, which can be used to face different diseases and problems”.

According to him, future work may find other markers from aerial images and, thus, refine these algorithms to ensure even greater reliability.

Drone or Satellite?

Although the aerial photos of Campinas were obtained with a drone, it is expected that, in the future, the strategy tested in this research will only use satellite images. “We used a drone because it was a study, but scanning with this equipment is expensive”, analyzes Chiaravalloti Neto.

“They also have less autonomy. To carry out a large-scale project, which includes large cities, we will need satellite images”, completes Santos. In the study in Belo Horizonte, satellite images were used successfully — they need high resolution for the computer to be able to identify the patterns. According to Santos, access to this type of image is fortunately expanding.

Although this type of methodology seems expensive, it generates potential savings by eliminating the need for face-to-face visits to map, house by house, areas susceptible to dengue. Instead, health agents would take advantage of information obtained remotely — and processed with artificial intelligence — to go to priority locations with more assertiveness.

Next steps

The current model is able to detect water tanks, but not if they are properly sealed. Something similar applies to swimming pools: he identifies them, but does not know if they are well maintained or closed. “This methodology could be refined to differentiate well-kept structures from those that would actually serve as breeding grounds for mosquitoes”, says Chiaravalloti Neto. Accusing these patterns and other structures linked to greater mosquito infestation would make the algorithm even more reliable for defining public health measures.

At the moment, researchers are installing a series of traps for the Aedes aegypti in about 200 blocks of Campinas and evaluating in detail the conditions of the properties and the presence of different mosquito breeding sites. Socioeconomic indicators will also be examined. The next step will be to evaluate aerial images of these regions with the same logic used in the aforementioned research to classify the degree of risk for the presence of the Aedes aegypti and the diseases transmitted by it.

“By observing these blocks, we intend to build a model for prioritizing dengue control for the entire city and, later, for the rest of Brazil”, concludes Chiaravalloti Neto.

In addition to funding from Fapesp, the researchers had resources from the Serrapilheira Institute, the CNPq (National Council for Scientific and Technological Development), the Dean of Research at USP and the Research Support Foundation of the State of Minas Gerais (Fapemig) . Sucen also provided structural support.

The authors involved are: Higor Souza Cunha, Brenda Santana Sclauser, Pedro Fonseca Wildemberg, Eduardo Augusto Militão Fernandes, Jefersson Alex dos Santos, Mariana de Oliveira Lage, Camila Lorenz, Gerson Laurindo Barbosa, José Alberto Quintanilha and Francisco Chiaravalloti-Neto.