Bruise detection in Pacific pink salmon (Oncorhynchus gorbuscha) by visible and short-wavelength near-infrared (SW-NIR) spectroscopy (600-1100 nm)

Visible and short-wavelength near-infrared (SW-NIR) spectroscopy (600-1100 nm) was used to detect bruises in intact, whole Pacific pink salmon (Oncorhynchus gorbuscha). The measurements were performed noninvasively through the skin and scales in the diffuse reflectance mode. Digital images of bruised and nonbruised regions of fish were captured after the fish samples were filleted. Image analysis was conducted using Adobe Photoshop 7.0 with relative gray values used as reference values in a partial least-squares (PLS) model. A PLS cross-validation model using six latent variables yielded a standard error of prediction (SEP = 0.05%, R = 0.83). Approximately 84% of all nonbruised spectra were correctly classified, whereas approximately 81% of all bruised spectra were correctly classified. These results suggest that visible and SW-NIR could be used to control the bruise defect of fish products during processing, thereby improving product consistency and quality.