The invention discloses a method for quickly and non-destructively detecting variety,
sugar degree and acidity of an apple, and belongs to the technical field of food detection. The method disclosed by the invention comprises the following steps: acquiring a near
infrared spectrum of an apple sample, and simultaneously, respectively detecting the
sugar degree and the acidity of the apple through a handheld
refraction sugar degree meter and a pen type
pH meter; then, performing dimension reduction
processing on the pre-processed apple spectrum data with a
principal component analysis method, and optimizing by using a
genetic algorithm; finally, classifying with a BP neural
network method, and finally, determining an optimal prediction model through repeated
verification. According to the BP neural network-based prediction model established by the invention, the variety, the sugar degree and the acidity of the apple can be predicted very well. In a conventional method, the apple is required to be damaged, so that the conventional method is destructive detection; moreover, the sugar degree and the acidity are required to be detected respectively; the variety of the apple cannot be detected through a suitable method. The method has the advantages of non-destructiveness, quickness, low cost, no sample pre-treatment, no chemical
reagent, no
pollution and the like; moreover, the variety, the sugar degree and the acidity of the apple can be predicted simultaneously.