Ispitivanje mogućnosti primene kompjuterskog vizuelnog sistema za merenje boje hrane animalnog porekla

Abstract

Boja predstavlja veoma važno senzorno svojstvo animalnih proizvoda, koje u velikoj meri utiče na izbor i odluku potrošača prilikom kupovine. Potrošači povezuju boju sa velikim brojem faktora kvaliteta, pa je stoga ona veoma bitan pokazatelj svežine, zrelosti i sveukupne prihvatljivosti hrane animalnog porekla. Zbog činjenice da predstavlja subjektivni doživljaj posmatrača, i da na njenu percepciju može da utiče širok spektar faktora, boju je neophodno objektivno izmeriti. Merenje boje hrane animalnog porekla je od fundamentalne važnosti, jer ona određuje pravu komercijalnu vrednost proizvoda. Objektivno merenje najčešće se postiže korišćenjem tradicionalnih uređaja - kolorimetara. Ova metoda za merenje boje je brza i jednostavna, a kalibracija se postiže primenom standardnih pločica na početku rada. Iako kolorimetri obezbeđuju jednostavnu i brzu procenu boje, često mere malu površinu uzorka, pa stoga nije moguće dobiti očitavanja cele površine uzorka u jednom merenju. Međutim, ova vrsta metode ima i nedostatak u pogledu ponovljivosti i tačnosti. Pored toga, hrana mora biti uniformne površine, kao i boje. Zbog ovih pomenutih nedostataka kolorimetra, razvijen je novi, alternativni metod poznat kao kompjuterski vizuelni sistem (CVS). Ovaj metod omogućava veću brzinu merenja, objektivnost i efikasnost, kao i trajno čuvanje podataka za kasniju analizu. Osnovni cilj ove doktorske disertacije bio je da se ispita mogućnost primene CVS metode za merenje boje hrane animalnog porekla. Drugi cilj istraživanja bio je utvrditi postojanje razlika u boji dobijenih pomoću kolorimetra i CVS-a. Dobijeni instrumentalni rezultati boje za hranu animalnog porekla pokazali su da je postojala statistički značajna razlika između dva uređaja, iako su sprovedeni isti uslovi merenja. Učestalost sličnosti (test A) pokazala je da su CVS-dobijeni čipovi u boji slični bojama uzoraka u svim ispitivanjima. U drugom testu (test B), CVS-dobijene boje bile su sličnije bojama uzoraka prikazanih na monitoru, u poređenju sa bojama dobijenim pomoću kolorimetra. Treći test (test trougla) pokazao je da postoji razlika između boja dobijenih pomoću ova dva uređaja. Naime, rezultati su pokazali da je moguće primeniti CVS za merenje boje hrane animalnog porekla. Takođe, dobijeni rezultati merenja bili su ponovljivi i reproduktivni. Na osnovu ovih rezultata utvrđeno je da je CVS pouzdan i da ga treba smatrati superiornim alatom za zamenu tradicionalnih uređaja.

Color represents quite significant sensory feature of animal source foods, which has a great influence on purchase decision of consumers. Customers connect it with a large number of quality factors and, consequently, it speaks on behalf of freshness, ripeness, and overall acceptability of animal source foods. Owing to the fact that the color represents the subjective experience of the observer, and its perception can be affected by a variety of factors, the color is necessary to be objectively measured. The color evaluation of animal source foods is of fundamental importance and acts in favor of determining the real commercial value of the product. The objective color analysis is commonly performed using traditional devices – colorimeters. This method is both fast and simple and calibration is reached via using the standard plate on the beginning of the process. Although, colorimeters provide a simple and rapid color assessment, often evaluate a single position, therefore, it is incapable to obtain the readings of the whole sample area in a single evaluation. However, this type of method also has a drawback in terms of repeatability and accuracy. Additionally, the food has to be uniform surface as well as color. On account of this mentioned drawbacks of the colorimeter, it is developed new, alternative method known as a computer vision system (CVS). This method offers rapidness, objectiveness and efficacy, as well as permanent storage of data for further analysis. The main aim of this doctoral dissertation was to examine the possibility of applying the CVS for the color assessment of animal source foods. The second goal of the research was to determine the existence of the color differences obtained using the colorimeter and the CVS. The obtained instrumental color results for animal source foods revealed the statistically significant difference, even if the same measurement conditions were performed. The frequency of similarity (test A) showed that CVS-generated color chips were similar to the color of samples in all trials. In the second test (test B), the CVS-obtained colors were found more similar to the color of samples visualized on the monitor, compared to colorimeter-obtained colors. The third test (triangle test) showed the difference between colors measured by these two devices. Namely, the results showed that it is possible to apply a CVS for measuring the color of animal source foods. Also, the obtained results of measurement were repeated and reproductive. According to these results, it can be concluded that the CVS is found to be reliable and should be considered as a superior tool for replacing traditional devices.