Increasing population and limitation of arable land, has required humans to produce as much food as possible through alternative sources such as mushrooms. Today, cultivation of edible mushrooms has become very important in the world. Mushrooms are relatively good sources of vitamins, proteins and amino acids. High-quality mushrooms have the appropriate color, moisture content and texture and a firm cap free of brown spots. Other than color, eva‎luation of other quality components of mushroom by using reference methods has a destructive nature and requires a lot of time and laboratory tasks. Therefore, it is very important to develop a method that can measure other quality components of mushroom in a non-destructive, accurate and fast way, in addition to color eva‎luation. In this study, two non-destructive methods of Vis/NIR spectroscopy and hyperspectral imaging (HSI) were used to eva‎luate the quality of edible mushrooms. The quality parameters eva‎luated in this study included firmness, moisture content and color indices including L *, a * and b *. For this purpose, 300 edible mushrooms were prepared from Yekta Sepahan Mushroom Production Company from different harvests. The experiments were performed in 5 different rounds, including the day of receiving the sample, the third, fifth, seventh and ninth days after production. After measuring physical characteristics (including weight and size) and color (L *, a * and b *), hyperspectral images of the samples were recorded using a reflective imaging system in the range of 450-950 nm. Then by using visible-near-infrared (Vis/NIR) spectrometer, the spectra of the samples were recorded in interactance mode in the range of 400-1100 nm. Immediately after recording hyperspectral and spectroscopy data, firmness and moisture content indices were measured by standard methods. To analyze hyperspectral images, first, for each sample, a wavelength image of 628 was selected and by image processing methods, its background was removed and finally, a binary image was obtained. By multiplying the binary image in the images obtained from other wavelengths, the hyperspectral image of that sample was obtained without background. Then the mean spectrum of each sample was calculated. The relative spectra of both HSI and Vis/NIR spectroscopy were then calculated and preprocessed by different methods. The partial least squares (PLS) method was used for regression analysis. The results of this study showed that, with the exception of the firmness, other important qualitative indicators of the mushrooms, including moisture content and color, can be successfully detected and predicted by using Vis/NIR and HSI spectroscopy methods. While the Vis/NIR spectroscopy method was able to predict the firmness (RPD = 1.444) and b * (RPD = 3.000) with a slight difference compared to the HSI method, HSI method outperformed the Vis/NIR spectroscopy in estimating moisture content (RPD = 2.298), L * (RPD = 3.088) and a * (RPD = 2.231). Considering that the Vis/NIR spectroscopy method was performed on a limited number of points of the mushroom cap (mean of 3-point spectrum), the superiority of the HSI method can be attributed to the ability of this method to record spectral data from a large number of points (pixels) belonging to the cap, which can describe the changes in the quality characteristics of the mushrooms better and more completely

احمد ميره اي

مرتضي صادقي، ميلاد فتحي