Abstract
Monitoring insects poses challenges due to the need for prolonged efforts to understand trends effectively. These efforts are often time-consuming, labor-intensive, and costly, requiring taxonomists for species identification. These constraints limit research and monitoring, particularly given that insects are underfunded in the field of conservation. Despite these challenges, insect monitoring and conservation are vital for protecting habitats and irreplaceable ecosystem services they provide. To address these limitations and enhance monitoring practices, approaches that greatly reduce the reliance on taxonomists and involve participatory science are crucial. The program, Measuring Insect Board Coverage using Adaptive Thresholding (MIBCAT), aims to achieve this by shifting the focus to computer vision and examining the overall composition of insect diversity and abundance rather than species identification. While the use of computer vision in the field of entomology is not new, most focus on automatic species identifications via machine learning; MIBCAT takes on a new approach. MIBCAT employs adaptive thresholding to find insect contours on a moth sheet (board) image. Multiple characteristics of each contour can be measured, including color and size variables. Hierarchical clustering categorizes these measurements into similar and distinct size and color groups, offering potential species sorting capabilities. Furthermore, by assessing the size of insects on the board, the total surface area coverage of the board is measured. This approach is comparable to assessing biomass, and provides a time-efficient, cost-effective, and accessible method for monitoring insect abundance over time. Despite the hierarchical clustering method needing major improvements before reaching similar capabilities as a taxonomist, the assessment of total board coverage proved notably successful in gauging overall insect abundance. With further refinements, this metric holds promise as a valuable tool for future monitoring, providing crucial insights into the health of the entomofauna of a research site.