Fruit sugar degree detection method and system

The invention provides a fruit sugar degree detection method and system. The method comprises the following steps: collecting near infrared spectrum data of a fruit to be detected; inputting the near infrared spectrum data into a sugar degree detection model so as to obtain a sugar degree predicted value corresponding to the near infrared spectrum data according to an output result of the sugar degree detection model; the sugar degree detection model being constructed on the basis of a trained convolutional auto-encoder by adopting a transfer learning method. According to the fruit sugar degree detection method and system provided by the invention, on the basis of a transfer learning method, the trained to-be-migrated auto-encoder is utilized to establish the convolutional auto-encoder model for extracting the apple near infrared spectrum data features, so that the apple near infrared spectrum data set is utilized to train the convolutional auto-encoder model. According to the method, the problem that deep le