Skip to main content
ARS Home » Southeast Area » Athens, Georgia » U.S. National Poultry Research Center » Quality and Safety Assessment Research Unit » Research » Publications at this Location » Publication #377151
Research Project: Assessment and Improvement of Poultry Meat, Egg, and Feed Quality

Location: Quality and Safety Assessment Research Unit

Title: Prediction of quality traits of chicken breast fillets by different spectral range of hyperspectral imaging

Author
item YANG, YI - China Agricultural University
item WANG, WEI - China Agricultural University
item Yoon, Seung-Chul
item Zhuang, Hong
item JIANG, HONGZHE - China Agricultural University
item JIA, BEIBEI - China Agricultural University

Submitted to: ASABE Annual International Meeting
Publication Type: Proceedings
Publication Acceptance Date: 6/15/2018
Publication Date: 8/8/2019
Citation: Yang, Y., Wang, W., Yoon, S.C., Zhuang, H., Jiang, H., Jia, B. 2019. Prediction of quality traits of chicken breast fillets by different spectral range of hyperspectral imaging. ASABE Annual International Meeting. Paper No. 1800828. https://doi.org/10.13031/aim.201800828.
DOI: https://doi.org/10.13031/aim.201800828

Interpretive Summary: Spectral ranges of hyperspectral imaging between 400 and 2,500 nm were studied to develop models for prediction of quality traits of chicken meat. Two hypespectral image cameras in the visible and near infrared range from 400 to 1,000 nm and the near-infrared range between 1,000 and 2,500 nm were used to scan 60 individual chicken breast fillets. We found that the range of visible and near infrared (400-900 nm) was more suitable for the prediction of color and pH quality traits, whereas near infrared (NIR, 1000-2500 nm) was more suitable for the prediction of drip loss, expressible fluid, and salt-induced water gain. The predictive power of moisture was not good in both two spectral ranges. The results demonstrated that different spectral ranges would be necessary for prediction of different quality traits in order to archive better prediction performance.

Technical Abstract: Different spectral ranges of hyperspectral imaging were used in this study for the prediction of quality traits of chicken meat. By comparing the prediction ability of partial least square regression (PLSR) models, it is concluded that the range of visible and near infrared (Vis-NIR, 400-900 nm) is more suitable for the prediction of L* (Rcv=0.93 and RMSEcv=1.59), a* (Rcv=0.89 and RMSEcv=0.38), b* (Rcv=0.86 and RMEcv=0.88), and pH values (Rcv=0.80 and RMEcv=0.15). And near infrared (NIR, 1000-2500 nm) is more suitable for the prediction of drip loss (Rcv=0.72 and RMEcv=0.83), expressible fluid (Rcv=0.57 and RMEcv=2.07), and salt-induced water gain (Rcv=0.72 and RMEcv=18.30). While the predictive abilities of moisture were not good in both two spectral ranges (Rcv < 0.30). Our results of this study demonstrated that different spectral range can be used in the prediction of different quality traits in order to archive better prediction performance.