摘要

By using apple mechanical properties index (maximum stress, compression yield stress, elastic modulus), this paper uses BP neural network to establish artificial neural network model of apple hardness, and predicts apple firmness. With the mechanical properties index test for input and apple hardness as output, based on the error analysis of the network parameters, this paper can determine network topology and trained BP neural network, finally produce the network model. After the simulation, it can verify the neural network through the 4 group of non sample data, the relative error between the prediction value and measured value is below 1%. The neural network model can uses mechanical property index to predict apple hardness, and it can satisfy the accuracy requirements in the engineering application of apple hardness.

全文