Automating Moth Morphometrics

Authors: Jennifer Slater, Joe Elias

Field of Study: Biological Sciences

Program Affiliation: Undergraduate Research Award

Faculty Mentors: Eliza Grames

Easel: 63

Timeslot: Afternoon

Abstract: This project focuses on developing a computer program that can identify moths based on physical traits, such as wing shape, color and size. Moths are highly diverse with over 15,000 recorded species in the family Noctuidae alone, yet many species are still remain undocumented. This research aims to create an image analysis pipeline that can process large numbers of image files and identify moth species based on measurable morphometric traits, with a calculated percent error. Images of moths collected from black lighting will include a color scale, for standardized measurement of the moth’s size and color. The pipeline, developed in R, isolates moths, extracting wing outlines through edge detection, and quantifies wing angle, size and color distribution. Unlike existing tools that rely heavily on reference libraries, my system will focus on measurable visual traits, making it more effective for processing large data sets for field research. This work aims to help further our understanding of moth diversity.