Prediction of nutritional content of black soldier fly (BSF) larvae using multispectral imaging

TY - ABST

T1 - Prediction of nutritional content of black soldier fly (BSF) larvae using multispectral imaging

AU - Gebreyesus, Grum

AU - Jensen, Kim

AU - Luca, Alexandru

PY - 2024

Y1 - 2024

N2 - The economic viability of BSF production for livestock and pet nutrition heavily relies on the nutritional composition of its larvae, notably protein and fat content, influenced by diet, harvest age, and genetics. Maintaining consistent product quality necessitates continuous monitoring across larvae developmental stages. Selective breeding also offers potential for enhancing BSF larvae’s nutritional profile, requiring phenotypic data from potential live candidates. Conventional chemometric methods, while accurate, are costly, time-consuming, and destructive, hindering their suitability for live larvae phenotyping. Hence, there’s a pressing need for rapid, non-destructive, and cost-effective techniques to evaluate the nutritional content at different stages of the BSF larvae. Recently, spectral imaging has emerged as a powerful tool for food quality assessment, capturing both spectral and spatial data simultaneously. This study explores the potential of multispectral imaging in predicting individual BSF larvae’s nutritional content. With 71 larvae sampled and imaged, prediction models, employing 19 wave bands as features via PLS, exhibited promising predictive performance for lipid mass and lipid percent. With reasonable correlation between measured and predicted parameters for a test set of 21 larvae, the study demonstrates the promise of multispectral imaging in predicting individual larvae’s nutritional content, offering avenues for future research in nutritional content prediction from live samples.

AB - The economic viability of BSF production for livestock and pet nutrition heavily relies on the nutritional composition of its larvae, notably protein and fat content, influenced by diet, harvest age, and genetics. Maintaining consistent product quality necessitates continuous monitoring across larvae developmental stages. Selective breeding also offers potential for enhancing BSF larvae’s nutritional profile, requiring phenotypic data from potential live candidates. Conventional chemometric methods, while accurate, are costly, time-consuming, and destructive, hindering their suitability for live larvae phenotyping. Hence, there’s a pressing need for rapid, non-destructive, and cost-effective techniques to evaluate the nutritional content at different stages of the BSF larvae. Recently, spectral imaging has emerged as a powerful tool for food quality assessment, capturing both spectral and spatial data simultaneously. This study explores the potential of multispectral imaging in predicting individual BSF larvae’s nutritional content. With 71 larvae sampled and imaged, prediction models, employing 19 wave bands as features via PLS, exhibited promising predictive performance for lipid mass and lipid percent. With reasonable correlation between measured and predicted parameters for a test set of 21 larvae, the study demonstrates the promise of multispectral imaging in predicting individual larvae’s nutritional content, offering avenues for future research in nutritional content prediction from live samples.

M3 - Conference abstract in proceedings

SN - 979-12-210-6769-9

T3 - EAAP Book of Abstracts

SP - 536

EP - 636

BT - Book of Abstracts of the 75th Annual Meeting of the European Federation of Animal Science

PB - EAAP

CY - Rom

T2 - 75th Annual Meeting<br/>of the European Federation of Animal Science

Y2 - 1 September 2024 through 5 September 2024

ER -