Nondestructive Inspection Of Melon's Sugar Content Based On Dielectric Properties

Non-destructive measurement and recognition of fruits' internal quality is apromising research topic in the area of agricultural and food engineering technology,which will be significant to meet the requirement of food quality and safety, toimprove the market value and market competition, and to increase the income of thefarmers. At present, conventional analytical methods used for fruits' internal qualityare destructive, time-consuming, expensive and low precision, and therefore it is animportant work to find non-destructive measurement method for fruits' internalquality based on dielectric properties and to build the mathematic models andevaluation system of stable performance.Presently, however, the estimate of melons' solid sugar content is still using thetraditional identification of calculating, watching, listening, measuring, smelling andother methods. These methods, depending on personal experience in most time, aredifficult to make accurate estimation of melons' solid sugar content; meanwhile, it is alarge scale-labor and time-consuming procedure. With the diversification of melons aswell as their cultivated forms, these methods can hardly meet the current needs ofproducing and marketing. Therefore, an accurate, rapid and simple non-destructivetesting of melons' solid sugar content technology has a great theoretical and practicalsignificance for grading, transportation, as well as sales links.Nondestructive quality measurement technique has advantage of acquiring thesample's internal quality index without ruining the samples. The paper pays attentionon the feature fruit melon of Xinjiang Province. It is a well tasted fruit by domesticand foreign consumers. The solid sugar content was chosen as the reference index.Areas including acquisition of melons' dielectric properties, extraction and analysis ofthe dielectric information, establishment and comparison of the mathematical models.On this basis, the model of non-destructive measurement of melons' internal quality, based on principle component analysis combined back propagation neural network(PCA-BPNN) optimized by genetic algorithm, had been established. The major workis summarized as follows:Firstly, the principles and methods of the non-destructive measurement based ondielectric properties were clarified. Then the measurement platform of melons'dielectric properties was established. The influences of measurement testing parts forthe signal response were explored in order to find the best conditions to get moreaccurate data of the melons.Secondly, methods of extraction for dielectric information were studied. Featureextraction depending on principal component analysis, the spectral feature dimensionswere reduced under the premise of retaining the useful information, which laid thefoundation for the future establishment of mathematical models.Thirdly, the linear and nonlinear mathematical models of melons' sugar contentwere established based on dielectric properties. The simulation results showed that theprediction accuracy of the linear model for melons' internal quality was not satisfied,while the non-linear back propagation artificial neural network model improved theprediction accuracy, however, there still existed the problems of unstable of thetraining process and shortage of generalization ability. The genetic algorithm wasapplied to optimize the parameters of weights and thresholds. This model obviouslyshowed the better predictive performance than the back propagation artificial neuralnetwork model.The results showed that the non-destructive method which adopted dielectricproperties utilizing back propagation optimized by genetic algorithm has a higherperformance on predicting the solid sugar content of melons. This laid a solidfoundation for the future research of melons' non-destructive detection.