Design and Implementation of Fruit and Vegetable Recognition System Based on Deep Learning

Abstract

With the development of agricultural intelligence, the market of fruits and vegetables continues to expand. In order to reduce the labor force in the processing and classification of fruits and vegetables, and enable individual users to identify fruits and vegetables at any time,
this paper designs and develops a fruit and vegetable identification system. Through the demand analysis of the system, the functional module
design and neural network model design were completed, and then the convolutional neural network model was built based on the PyTorch
framework, the fruit and vegetable dataset was loaded, the training model improved the accuracy, and finally the well trained model was called
to realize the recognition of fruit and vegetable maps. The system realizes the user interface, which is convenient for users to call the model to
realize the recognition of fruits and vegetables. After testing, the system has stable performance, basically meets the needs of users, and realizes the expected functions and effects, which has certain practical value.