Quality and shelf-life prediction of cauliflower under modified atmosphere packaging by using artificial neural networks and image processing

Abstract The aim of this research is to study the effects of modified atmosphere packaging (MAP) with a gas mixture (92% nitrogen, 5% carbon dioxide and 3% oxygen) and a packaging with ordinary air on shelf-life and the quality of cauliflower after harvesting. Polyethylene (PE) and polypropylene (PP) pouches with 40 µm thickness were used for packaging. Four types of packaging were studied; cauliflower packaging using PE pouches with ordinary air (AP 1 ), cauliflower packaging using PP pouches with ordinary air (AP 2 ), cauliflower packaging using PE pouches with MAP (MAP1), and cauliflower packaging using PP pouches with MAP (MAP2). These packages were stored for 30 days at a temperature of 4 ± 1 °C and relative humidity of 90 ± 5% in a refrigerator. The results showed that cauliflower packed with AP 1 , AP 2 and MAP2 methods had a shelf-life of up to 30, 20 and 25 days, respectively. While, the shelf-life of the cauliflower packed with MAP1 could be increased to more than 30 days. In order to predict the maximum shelf-life of crop packaged with MAP1 method based on the acceptable maximum color changes, various multi-layer perceptron (MLP) were designed and evaluated. According to the analysis and the comparison of the mean square error (MSE) and the coefficient of determination (R 2 ) related to the test data, the best artificial neural networks (ANN) model was obtained from the MLP with one hidden layer of 12 neurons. MSE and R 2 of the optimum network were 0.0095 and 0.990, respectively. The results of ANN showed that in terms of total color changes, cauliflower crop packed with MAP1 type had marketing capability of up to 50 days. The results of statistical analysis showed that mechanical properties changes among packaging methods on days 20, 25 and 30 had no significant difference. While differences in color changes and weight loss had significant difference in these comparisons.