1. Institute of Remote Sensing and Digital Earth, Chinese Academy of Sciences, Beijing 100101, China
2. Key Lab of Spatial Data Mining & Information Sharing, Ministry of Education, Fuzhou University, Fuzhou 350000, China
3. College of Resources and Environmental Sciences, Hunan Normal University, Changsha 410081, China
4. University of Chinese Academy of Sciences, Beijing 100049, China
Abstract：While it is easy to rotten resulting from numerous factors including the enzyme reaction and microbial reaction during its storage, transportation and varied fine processes, pork is one of the main meat products daily consumed in China, . Consequently, before brought onto the dinner table, the pork bought from the market may be not fresh, which will not only affect the taste but also human health. Pork freshness is considered as one of the major driving factors for consumers to buy or not. Therefore, it is very important and also urgently needed to detect pork freshness in situ timely, quickly and accurately in order for consumer’s food safety. Compared to traditional physical and chemical testing methods operated in laboratory, the novel visible/near-infrared spectral analysis technology for detection of cold fresh pork quality has attracted wide interest recently, due to -features quickness, high efficiency, non-destruction and non-contact of the visible/near-infrared spectral analysis technology.And hence it is more suitable for food safety quick-detection. However, most of current researches have focused on the spectral model development based on statistical methods, which results in lack of physical meaning and poor applicability, and hence hinders the application and popularization of the spectroscopy technology. The visible/near-infrared spectral features of pork with varied freshness were investigated in this paper. On this basis, a pork freshness spectral feature index (FI) was constructed using the stable absorption property of pork myoglobin at 760 nm. Additionally, the sensitivities of FI to spectral resolution and signal-to-noise (SNR) were analyzed by simulations. Our research indicated that FI was simple but had a clear physical meaning, and can become a good proxy of pork freshness. Furthermore, the FI showed a relative low dependency on the spectrometer’s key properties, such as spectral resolution and SNR. It may work well as long as the spectral resolutions at both 760 nm and adjacent bands are better than 10 nm, and the SNRs are no lower than 45∶1. This study may provide a scientific basis for the design and development of low-cost and handheld portable spectrometers aiming for pork freshness quick detection in consumer markets.
