This report reviews various measurement techniques and methods for assessing meat quality in lamb and beef after slaughter. It has mainly been based upon searches in databases of scientific literature, but also on discussions with industry people and colleagues.
The concept of meat quality is multifaceted and not easily defined. At the same time, most people who appreciate a good piece of meat would agree on that tough meat should be avoided. Hence, one property that has attained considerable attention in the literature is tenderness. Another characteristic that is usually emphasized as something positive is marbling, partly due to several studies that have indicated a relation between marbling and taste, juiciness and tenderness, all being properties that are important for a positive eating experience. Other features that attract attention are e.g. fatty acid composition, water retention ability, pH and colour.
The first attempts to find an objective method for tenderness measurement were based on mechanical methods. Essentially a spear-like object that was shot into the meat, while the resulting force was measured. For measurement of other parameters, such as pH, relatively traditional techniques were similarly used, based on more or less analogue technology. However, in recent decades there has been an increasing amount of studies using techniques that have benefited from the exponential development of digital and solid-state technologies. This development has e.g. led to easier ways to generate, measure and analyse electromagnetic, optical and acoustic signals.
A common approach in simpler measurement methods is trying to find an algorithm that is based on analysis of the frequency response of a signal that typically may be of electrical nature, ultrasound or light. Methods utilizing NIR have been particularly promising. One example of NIR equipment is NitFomTM, which is used for quality assessment of fat in pork meat. Methods based on measurement of electrical impedance have also, at least periodically, found establishment on the market.
There are also more advanced approaches, with the ambition to obtain spatial resolution of properties within the object under assessment. For measurements with three-dimensional resolution, primarily computer tomography imaging (CTI) and magnetic resonance imaging (MRI) are candidates, but also ultrasound. Hitachi-Aloka, makes ultrasonic equipment that can be used for scanning live animals, and another example of ultrasonic equipment is ECM EXAGO. MRI and CTI, however, are still too expensive, advanced and slow to be realistic alternatives for online measurement in the industry in the near future.
MRI and CTI are developments of NMR (Nuclear Magnetic Resonance) and X-ray. While also NMR seems to be a little too expensive for industrial applications, and primarily a lab tool, X-rays have been used in the meat industry since the 1970s, e.g. for measurement of fat content. One commercial product that early found establishment on the market is Anyl-Ray Oystar.
The development of the digital camera opened up for advanced image processing. There are several studies based on analysis of the kind of information that can be extracted from RGB images, so-called vision technology, but also on analysis of images that contain much more detailed spectral information, so-called hyper- or multispectral image analysis. However, analysis of the amount of information that is collected with the latter kind of technology requires large computation and data management capabilities. In this context, the continuing development towards more accessible computational power is highly valued.
While image processing based on RGB information has been successful in measuring more or less what is also perceived by human eye (in this context such parameters as marbling and colour), hyperspectral image analysis has shown potential to go one step further. In addition to visual properties, the technology has shown promising results in measuring such things as chemical composition (e.g. proportions of fat, protein and water), pH and tenderness. Much resources have been invested in development of functional systems for online classification of meat in the industry. The results have been promising, and companies have been started up for the purpose, but the definitive breakthrough has not yet taken place.
In conclusion, several attempts have been made to find objective measurement methods for assessing and potentially classifying meat quality. Many promising results have been reported in the literature. Yet it is difficult to make any recommendations on one single salvaging technique based upon these results. Possibly, the technology that is currently attracting the most attention and hopes is hyperspectral image analysis, especially if the intention is to find a technology, suitable for forming the basis for a classification system. In such a context, hyperspectral imaging is a technology that meets many positive criteria: it is contact-free, it has spatial resolution, it combines advantages of both vision and NIR. There are also several studies that show promising results, and there is still good hope that the technology will develop further in near-time (both in terms of price and performance), hand in hand with the trend in society towards increased digitalisation (i.e. development of artificial intelligence, better and cheaper sensors, increased access to computational power, connected devices, etc.).