Prediction of broiler shear force using near infrared spectroscopy with second derivative linear modeling

Abstract

This study explores the use of linear predictive models, specifically principal component regression (PCR) and partial least squares (PLS), in combination with a cost-effective near infrared spectroscopy (NIRS) system to noninvasively assess the texture of raw broiler meat. The findings demonstrate that appropriate pre-processing techniques, such as excluding the visible spectrum and applying the second-order Savitzky-Golay (SG) derivative with an optimal filter length (FL), enhance model performance. Notably, the PLS model outperformed PCR, requiring fewer latent variables (LVs) to achieve accurate predictions. This suggests that PLS more effectively captures key spectral features associated with meat texture, making it a promising approach for assessing raw broiler meat quality in a practical, cost-efficient, and non-invasive manner. These results highlight the potential of integrating linear predictive models with NIRS technology for reliable texture analysis in the poultry industry.