As global temperatures rise, dengue-carrying mosquitoes are expanding into new regions, including parts of the southern United States and Europe. Diseases such as dengue, chikungunya, and Zika spread through mosquitoes that breed in stagnant water found in shallow containers like discarded tires, flower pots, or small pools.
Consequently, eliminating these breeding sites has become a critical strategy in preventing outbreaks. However, identifying them is often difficult. Many breeding locations—such as old tires hidden behind fences or covered by vegetation—remain unnoticed during traditional ground surveys.
Stanford Researchers Introduce AI and Drone-Based Surveillance
To address this challenge, researchers at Stanford University have developed an innovative approach that combines drones with artificial intelligence (AI) to detect hidden mosquito breeding sites quickly and accurately. Their findings, published in the journal Remote Sensing Applications: Society and Environment, demonstrate how advanced technology can transform mosquito surveillance and disease prevention efforts.
According to Joelle Rosser, Assistant Professor of Medicine–Infectious Diseases at Stanford Medicine, recent advances in high-resolution remote sensing and AI have enabled scientists to rethink how they study the interaction between environmental changes and human health. She noted that these technologies offer powerful tools to respond to the rapidly evolving challenges posed by climate change.
Identifying Hidden Breeding Sites in Urban Landscapes
The research team tested their method in Makassar, Indonesia, a densely populated coastal city known as a dengue hotspot. Across the city, thousands of discarded tires lie scattered on rooftops, hidden behind fences, or concealed by vegetation. These tires collect rainwater and create ideal breeding environments for Aedes aegypti, the mosquito species responsible for transmitting dengue and other viral infections.
Because these breeding sites are often hidden across large urban areas, public health workers face significant difficulties locating and eliminating them through conventional surveys.
Drone Imaging and AI Algorithms Improve Detection
To overcome these limitations, the researchers collaborated with local drone pilots to capture high-resolution aerial images across a four-kilometre section of the city. They then trained two deep-learning AI systems to detect discarded tires by analysing their shapes, locations, and distribution patterns.
The results proved highly promising. Both AI models—based on advanced convolutional neural networks—demonstrated high accuracy. Notably, one model called U-Net++ identified nearly twice as many discarded tires as human analysts reviewing the same images. The AI systems even detected tires partially submerged in water or hidden beneath plants and shadows.
Expanding Surveillance to Hard-to-Reach Areas
Researchers highlighted another key advantage of drone technology: its ability to reach remote or inaccessible locations. According to Mehedy Hassan, a postdoctoral researcher at Stanford’s Department of Medicine–Infectious Diseases and co-first author of the study, drones can significantly expand surveillance coverage while reducing operational costs.
When combined with AI-powered detection, drones enable rapid identification of high-risk mosquito breeding sites. As a result, public health teams can focus their efforts on targeted interventions rather than conducting time-consuming manual inspections.
A Timely Strategy for Disease Prevention
As reported by medicalxpress, the researchers emphasised that this technological approach comes at a critical time. Climate change has increased temperatures and intensified rainfall and flooding in many regions, creating favourable conditions for mosquito populations to grow.
Moreover, with limited vaccines and no specific treatments available for many mosquito-borne diseases, controlling mosquito breeding remains one of the most effective strategies for disease prevention. Rosser noted that rapid identification of high-risk breeding sites could significantly strengthen vector control programmes.
Future Research and Global Applications
Looking ahead, the research team plans to expand their surveillance system to detect other types of waste that can collect water and support mosquito breeding. They also aim to explore multispectral drone imagery, which may further improve predictions about high-risk locations.
Importantly, the scientists intend to test whether this approach can be adapted to different geographical settings worldwide. Since Aedes aegypti mosquitoes already exist in the United States, the researchers believe the technology could help strengthen mosquito surveillance programmes in many countries.
Drones Emerging as Key Tools in Public Health
Beyond mosquito control, drones are increasingly proving useful in addressing broader environmental health challenges. In a separate study published in BMC Global and Public Health, Rosser and her colleagues outlined how drones can support research at the intersection of environmental change and human health.
Drones can quickly access hard-to-reach or hazardous locations, providing valuable visual data during disasters such as hurricanes. Emergency responders can also use them to assess damage and injuries rapidly. Additionally, organisations are already using drones to deliver medical supplies to remote or disaster-affected communities.
Monitoring Environmental Stress and Health Risks
Furthermore, drones offer a cost-effective alternative to traditional aircraft or satellite imaging. For example, farmers can use them to monitor crops for drought or heat stress during extreme weather conditions.
Rosser concluded that as environmental pressures intensify globally, drones will play an increasingly important role in monitoring and mitigating climate-related health risks. By providing rapid and precise environmental surveillance, these technologies can help public health officials stay ahead of emerging disease threats.