| Potato is one of the most important food and vegetable economic crops in the world, which has higher nutritive value and extensive use. Influenced by western diet culture and growth in the living standard, potato chips are currently fashionable as one kind of essential snack food in people’s leisure time. The potato cultivar’s quality is directly related to processing quality and economic benefits of the potato industry. Chinese potato industry possesses a wide variety and adequate supply currently. Quality indexes were different among cultivars, and these differences determine the quality of fried potato chips to a large extent. In this paper, NIR rapid detection of main potato’s indicators and suitability evaluation for potato chips processing were studied. The main contents are as follows:Firstly,200potato cultivars were collected. All samples were grown in the same place with the same fertilization. Potato harvest, store, processing and index analysis were conducted in the shortest time.8main quality indicators of potato (water, starch, reducing sugar, total sugar, ash, total soluble solid, Vitamin C and protein) were measured in laboratory.Secondly, NIR and PLS methods were used to establish4indicators’ prediction model respectively with fresh potatoes and potatoes after vacuum freeze drying. Results show that, vacuum freeze drying pretreatment of potato samples can effectively realize multi-component content prediction for potato.With all200different potato cultivars to augment the model, vacuum freeze-drying technique was used for sample pretreatment, mathematic models were established for prediction of4main processing quality indicators (water, reducing sugar, starch and protein) based on Fourier transform near-infrared spectroscopy (NIRS) technique and partial least square (PLS) method. The accuracy of models was estimated by the determination coefficients (R2), the root mean square errors of cross validation (RMSEC), the root mean square errors of prediction (RMSEP) and relative predictive determination (RPD). At the same time, external validation was conducted with50unknown samples. Results show that, the models were proven feasible to predict the main processing quality indicators of potato. Models of water and starch were effective which can be used in practical tests, however, precision of protein and reducing sugar models need to be further improved.Thirdly, In order to determine suitability of different cultivars for potato chips processing,74selected potato cultivars were investigated.8main quality indicators of potato (water, starch, reducing sugar, total sugar, ash, total soluble solid, Vitamin C and protein) and4quality indicators of potato chips (protein, sensory evaluation, crispness and whiteness) were measured in laboratory.56samples were removed randomly for the validation set, and other18samples for the calibration set. Comprehensive evaluation indicator of potato chips was calculated by normalization method and Euclidean distance of4individual indicators. Correlation analysis showed that the comprehensive evaluation indicator was significant to represent4individual indicators of potato chips. Principal component analysis was applied to select4irrelevant principal components (F1～F4) which could cover most information of potato samples. Stepwise regression method was used to establish the model of comprehensive evaluation indicator and4principal components (F1～F4). Substituted with original8main quality indicators of potatoes, the final model was determined. The determination coefficient R2is0.607, adjusted R2is0.585, F-value is26.815greater than F (0.001,3,52), and sig. value is0.000less than0.001. The model effect was relatively good with a high degree of fitting and significant correlation.18samples from the validation set were used to validate the effect of the model. The results show that the correlation coefficient for the measured value and prediction value of comprehensive evaluation indicator is0.502, sig. value is0.034which is less than0.05, and the model is applicable to actual suitability evaluation of potato chips processing.At last, Suitability of74cultivars for potato chips processing was preliminarily classified into3types based on K-means clustering method.15samples were selected as optimal cultivars for potato chips processing, which accorded with practical applications. However, the precision of the model need to be improved in future research. The results of this research can be applied in practice. Comprehensive evaluation indicator of potato chips can be determined without tedious laboratory analysis, proper potato material can be founded for chips’processing, and the result has guiding significance for chip processing industry. |
PDF Full Text Request