การตรวจสอบอาการเนื้อโพรงในผลมะม่วงพันธุ์น้ำดอกไม้สีทองด้วยเนียร์อินฟราเรดสเปกโทรสโกปี

Abstract

'Namdokmai Sithong' mango has the characteristic needs of foreign consumers. It is produced and exported in large quantities. Exporting fresh mangoes often have problems with poor quality yield, short shelf life, and other damaging factors. One cause of low quality is spongy tissue, which is a physiological disorder that cannot be seen from the outside and affects the quality of the fruit, economic value, and consumer confidence. Currently, near infrared spectroscopy (NIRS) is used to evaluate the product quality without destroying the samples (nondestructive testing). It is used to detect abnormalities occurring within a variety of agricultural products. This research aimed to use near-infrared spectroscopy to examine the spongy tissue of 'Namdokmai Sithong' mango fruit. The NIRS spectral wave number of 12500 - 4000 cm-1 was used to build a model. The model used stepwise discriminant analysis (SDA), partial least squares linear regression (PLSR), and artificial neural networks (ANN) to examine the spongy tissue in mango fruit cv. Namdokmai Sithong. According to stepwise discriminant analysis (SDA) was found that the training and validating data of first derivative spectra can classify normal and spongy tissue with an accuracy of 85.79 and 80.25 percent, respectively. The model used the partial least square (PLS) regression method was found that the training and validating data of the original spectra, first derivative spectra, second derivative spectra, standard normal variate (SNV) spectra, and multiplicative scatter correction (MSC) spectra had a low coefficient of determination (R2) and root mean square of error (RMSE). Artificial neural network (ANN) of short wavenumber near-infrared (12500 - 9000 cm-1) modified with first derivatives. It showed that the training and validating data predicted normal tissue, with an accuracy of 99.24 and 98.98 percent, respectively. The training and validating data predicted spongy tissue, with an accuracy of 61.54 and 73.68 percent, respectively. Physical and chemical quality analysis revealed that the pulp color of the mango with spongy tissue was brown to black. The dry weight and total phenolic compounds of the spongy tissue were high. The total soluble solids (TSS) and titratable acidity (TA) were low. In the analysis of the physical and chemical quality, it was found that flesh color, texture, dry weight, TSS, TA, TSS:TA ratio, vitamin C content, carotenoid content, and total phenolic compounds can be analyzed by principal component analysis which was separated. The short-wavelength NIR spectrum can be used as a non-destructive quality evaluation of spongy tissue in 'Namdokmai Sithong' mango. The artificial neural network (ANN) model also provided 60.26 and 73.68% accuracy for predicting spongy tissues, respectively. Moreover, this research could be used as a guideline for developing methods to investigate other physiological disorders of 'Namdokmai Sithong' mango and other fruits.