Hyperspectral imaging and chemometrics assessment of intramuscular fat in pork Longissimus thoracic et lumborum primal cut

Abstract

To assess the intramuscular fat (IMF) of Longissimus thoracis et lumborum (LTL) primal cut excised from an animal carcass, rib sections are dissected for chemical or subjective evaluation. Dissection destroys the integrity of the LTL. This study examined the feasibility of developing a method based on hyperspectral imaging to predict the IMF of pork LTL using hyperspectral imaging. Spectral information from different tissues was collected at the cranial (Serratus ventralis, Spinalis dorsi, longissimus , backfat), and the caudal ( Gluteus medius, Psoas major ) ends of the LTL. The spectral information and the IMF content chemically extracted from the mid-loin (3rd and 4th last rib of the LTL) were used to develop regression models for the prediction of the IMF content in the 3rd and 4th. The individual results of the analysis from the various tissues at both the cranial and caudal ends show that models exhibited performances for IMF prediction with root-mean-square-error (RMSEP) ranging from 0.36 to 0.92% and coefficient of determination in prediction ranging from 0.43 to 0.81. Gluteus medius showed the highest accuracy. This study has demonstrated that hyperspectral imaging could be used to accurately predict the IMF content in the mid-loin by using spectral data from Gluteus medius . The result could be useful in segregating pork LTL based on IMF content. • Intramuscular fat in whole pork Longissimus thoracic et lumborum was evaluated. • Muscles from the cranial and caudal of the Longissimus thoracic et lumborum were scanned. • PLSR was used to develop models between tissues and intramuscular fat at 3rd/4th last ribs. • The Gluteus medius muscle predicted intramuscular fat accurately.