The aim of this project is to employ robotics, image analysis, and machine learning to automate the filtering rearing system (FRS) for the black pupal strain of mexican fruit fly to identify and…

The Animal and Plant Health Inspection Service awarded a Cooperative Agreement to the University of Hawaii at Hilo for $564,462 on September 15, 2025, under the Plant and Animal Disease, Pest Control, and Animal Care program (CFDA 10.025).

The award funds development of an automated filtering rearing system to remove recombinant Mexican fruit fly individuals from laboratory populations, preserving genetic translocations that link pupal color with fly sex. Deliverables include an imaging system with machine learning software to classify pupae by color and adult flies by sex; a database structure correlating pupal color classifications with adult fly sex; a carriage system carrying camera, microcomputer, and compressed air nozzle to destroy recombinant individuals; repurposing of a robotic gantry to scan and locate grids of pupae and flies; procurement of a pill shaker to facilitate pupal placement; and repurposing of a robotic loading and unloading system for grid storage and movement. Year 1 deliverable is an optical system capable of recognizing and making decisions based on pupal color and adult fly sex. Year 2 deliverable is a fully operational automated filtering system integrating the optical system with robotic loading and unloading capability, designed to operate with minimal human interaction and outside normal working hours.

Work is performed in Hilo, Hawaii with a period of performance extending through September 29, 2028. The system will support the APHIS Mexican Fruit Fly Program, the MexFly mass rearing facility, and agricultural growers.

The aim of this project is to employ robotics, image analysis, and machine learning to automate the filtering rearing system (FRS) for the black pupal strain of mexican fruit fly to identify and remove recombinant individuals from the rearing population and thereby ensuring retention of the genetic translocation that links pupal color with fly SEX. Activities will include: 1) development of an imaging system and machine learning software to classify pupae by color and adult flies by SEX; 2) development of a database structure to enable correlation of pupal color classifications with adult fly SEX; 3) development of a carriage system to carry the camera, microcomputer for image processing, and compressed air nozzle for use in destroying recombinant individuals; 4) repurposing of a robotic gantry to control movement of the carriage so grids of pupae and flies can be scanned/located; 5) procurement of a pill shaker to facilitate placement of pupae within grids; 6) repurposing of a robotic loading/unloading system to store grids in racks for easy movement and storage. This project will deliver an automated filtering system designed to work with minimal human interaction, perhaps even outside of normal working hours, so workers can establish rearing cages using populations of flies free of recombinants. The steps to produce such populations will be as shown in figure 5. At the end of year 1, this project will deliver an optical system capable of recognizing, and making decisions based on, pupal color and adult fly SEX. At the end of year 2, this optical system will be incorporated with a robotic loading and unloading system to make the overall system as hands-free as possible. Year 1: delivery of optical system. Year 2: operational AFS. This project will benefit the PPQ mexican fruit fly program, the mexfly mass rearing facility, and growers.