Automated computer vision monitoring of black soldier fly larval growth using metaheuristic optimization for waste-to-feed bioconversion
Journal
Computers and Electronics in Agriculture
Journal Volume
248
Start Page
111679
ISSN
01681699
Date Issued
2026-07-01
Author(s)
Abstract
Sustainable conversion of organic byproducts into agricultural resources is vital for achieving circular bioeconomy goals. The black soldier fly (BSF, Hermetia illucens) offers an efficient bioconversion pathway, transforming organic residues into protein-rich feed and biofertilizer. However, variability in feed composition and high labor requirements constrain the scalability and profitability of BSF-based waste-to-feed systems. This study introduces an automated computer vision framework for monitoring the growth of black soldier fly larvae (BSFL) using instance segmentation and metaheuristic optimization. Sewage sludge from Taiwan's wastewater treatment plants was used as a feed substrate, and a Pilgrimage Walk Optimization (PWO) algorithm was integrated with a You Only Look Once version 11-instance segmentation model (YOLO11-seg) to enhance hyperparameter tuning and detection accuracy. The model achieved high precision in estimating larval surface area and converting it to biomass weight, enabling real-time, non-contact monitoring of growth performance. An economic analysis using Net Present Value (NPV) and Internal Rate of Return (IRR) demonstrated that a 1:1 sludge-to-wheat bran mixture significantly improved bioconversion efficiency and financial returns, underscoring the feasibility of AI-driven computer vision for intelligent insect farming, organic waste valorization, and sustainable waste management. Nevertheless, these findings are based on laboratory-scale experiments, and field-scale studies are needed to validate scalability and real-world implementation.
Subjects
Bio-inspired optimization
Black Soldier Fly (BSF)
Computer vision
Economic feasibility
Instance segmentation
Larval growth monitoring
Sludge treatment
Waste-to-feed bioconversion
Publisher
Elsevier B.V.
Type
journal article