Recently Vice President for UF/IFAS, Dr. J. Scott Angle, introduced some of the Artificial Intelligence (AI) initiatives being implemented at UF/IFAS Tropical Research & Education Center (TREC). Among the initiatives he mentioned was Dr. Daniel Carrillo’s project to develop an automated trap to detect the presence of non-native fruit flies. Dr. Carrillo, an Associate Professor of Entomology and Nematology at UF/IFAS TREC, specializes in the biological control, ecology, and management of invasive insects and mites.
In Fall 2019, Dr. Carrillo partnered with Bioverse Labs to teach a machine to recognize non-native fruit flies in Florida. The new technology would, in theory, take an image of the sticky card on a fruit fly trap, scan the sticky card for insects present, and alert a scientist if any unusual or non-native fruit flies are identified.
While the task seems easy enough, training a computer to distinguish a fruit fly from other insects isn’t as easy as it sounds. Presently, Dr. Carrillo is generating large amounts of data necessary to train the automated trap system; this includes taking thousands of images of different fruit fly species. He expects that the technology may take another 2 years to fully develop the capabilities required to distinguish native from non-native fruit flies.