INVESTIGATION OF MEAT OPTICAL CHARACTERISTICS WITH THE PURPOSE OF IMPLEMENTING THE VISUAL METHOD OF CONTROLLING IT

Abstract

Meat and meat products should be included into the daily ration of individuals, therefore it is important to ensure high quality of meat products. Not only does adulteration of meat products reduce the quality of finished commodity but may constitute a dangerous factor for consumers’ health as well. The methods of controlling meat quality applied nowadays are characterized by a number of drawbacks – optional effect, expensive equipment, long examination period, the need to apply a great number of reagents, the need for competent staff and specialized laboratories. It is next to impossible to efficiently trace adulterated foodstuff using traditional methods. The methods currently in use are imperfect, labor-intensive, long in time and not sufficiently informative. That’s why it is important to develop effective meat quality control methods which would be accessible and affordable for the consumer and which would provide the opportunity not to conduct additional laboratory testing to get the assessment of the sample under study. The use of effective analysis methods should primarily ensure detection of adulterated products and determination of the meat freshness degree.

Different types of meat are represented in the meat market being in permanent demand with consumers, and sometimes consumers have problems with choosing a high-quality product out of all this variety. Currently, physical and chemical methods are mainly used to trace adulterated meat products, and they are applied with the help of physical devices in laboratory conditions, viz.: spectral, electrochemical, electrophoresis and luminescent methods. All those methods are characterized by a number of drawbacks, among which are: optional effect, expensive equipment, high time-intensity and labor-intensity, the need to apply a great number of reagents, the need for competent staff and specialized laboratories. Thus, application of physical and chemical methods to identify types of meat is hindered due to their complexity and large time span necessary to do the measurements.

Therefore, currently, one of the topical tasks in the practice of consumer rights protection is development of new and improvement of available methods of controlling products of animal origin. The main requirements to modern methods of controlling the quality of animal-derived products can be formulated as follows:

• eligibility for effective identification of meat product types;
• identification period should not create obstacles for timely acquisition of meat products by consumers;
• identification likelihood should correspond to the acceptable consumer risk.

Thus, optical method of efficient meat quality identification is suggested, its essence being as follows:

• development of virtual optical benchmarks of different meat types;
• determination of the characteristic features of each specific meat type on the basis of photo images of different meat types;
• identification of the level of meat quality based on the theory of image recognition.

The implementation algorithm has been developed to implement the suggested method, and it constitutes the basis for the development of the respective software that may be installed, let us say, on consumers’ mobile phones.

This approach will enable to develop meat product identification methods, improve efficiency and effectiveness of metrological performance in the field of meat products manufacturing and distribution.

To implement the suggested method it is necessary to study the terms of its application and to develop a mathematical model of optical meat identification. With this in view the main factors of influence have been determined, viz.: background (image visibility) and illumination level of the sample under study.

The images were identified through analysis of red (RED), green (GREEN) and blue (BLUE) spectra in the obtained photo images. With a view to developing the benchmark base for determining the quality of chicken meat 25 samples were examined. And it was established that at the illumination level L>1600 lx this regularity (fig. 1) disappears, therefore the recommended value for illumination L should not exceed 1,600 lx during the implementation of the suggested method of meat quality identification.

The article presents a formalized mathematical model of optical identification of meat types and suggests the model of the critical rule for decision-making purposes.

It has been established that the mathematical model is adequate within the range from 730 lx to 1,850 lx.

Similar studies of the researched object were also made against other backgrounds, however no stable and regular dependences have been traced. Therefore, it is recommended to implement the method against the white object background.