Florida is experiencing one of its most serious dengue outbreaks in the past century. Against this backdrop, the startup Digitomy, founded at the University of South Florida (USF), has developed a mosquito trap with artificial intelligence (AI). In laboratory tests, the accuracy of species identification exceeded 98%.
Research at USF has been underway since 2018. The technology was created by associate professor Ryan Carney and professor Sriram Chellappan. The system uses AI to determine a mosquito's species, its sex and whether it has fed on blood. The data are transmitted to the cloud (remote servers for storing and processing information) and displayed on a map. This is intended to help health authorities monitor disease vectors and carry out more targeted treatments.
"The mosquito is the deadliest animal on the planet. Although there are 3,500 species of mosquitoes, dengue in Florida is spread mainly by one - Aedes aegypti. Data from our traps allow us to build a habitat model and track in real time the presence and abundance of this species in order to purposefully curb its spread," Carney was quoted as saying by the USF press service.
The trap is hung on a low tree branch or placed on the ground. A light or bait on top attracts mosquitoes, and a fan draws them inside. The insects stick to an adhesive pad, after which two built-in cameras photograph them, and the AI compares the images with a large database of mosquito images and identifies the species.
"Traditional monitoring requires manual work by specialists and can take two to three days under a microscope. The AI in our trap speeds up the process and allows targeted measures to be taken faster, while the decision remains with a human," Chellappan explained.
Laboratory calculations showed that recognizing all the mosquitoes on a single adhesive pad requires less energy than publishing one Instagram post. The technology was tested at USF and in partnership with the Anastasia Mosquito Control District in St. Johns County, including the city of St. Augustine. "The collaboration with Digitomy and USF will help us implement new AI technologies. We will be able to better monitor vector mosquito populations and engage the public in surveillance," said Rui-De Xue, executive director of the district.
In addition to identifying species, the system can indicate where disease vectors are concentrated and complement the DYCAST early warning system for dengue and West Nile fever outbreaks, created by Carney earlier. Trap data are also added to a closed layer of the Global Mosquito Observations Dashboard, which has already collected more than half a million reports from volunteers. In the future, this will help control services respond faster to outbreaks.
In addition, Digitomy is developing a cloud-based educational platform: users will be able to study mosquito images, identify species and receive feedback from the AI without attending in-person classes. "Education is our second line of defense. Traps provide data at the macro level, and people should remember three rules: drain standing water, cover skin with clothing and use repellent with DEET," Carney noted.
The developers are now preparing for field trials in Hillsborough County and are simultaneously working with the USF technology transfer office to license the system to trap manufacturers. No timeline for release to the consumer market or price was reported.