Research On The Recognition Of Fruit And Vegetable Based On Embedded System

With the improvement of people’s quality of life,the consumption level of fruits and vegetables has increased sharply,resulting in a increase in sales and an increase in workload.Therefore,using artificial intelligence technology in saling fruits and vegetables to reduce the workload and improve the work efficiency.At the same time,considering the convenience and quickness of the application of embedded system in daily life,this paper designs a set of embedded fruit and vegetable image recognition system based on Android system platform,which can be used in daily life in the sales of fruits and vegetables.Aiming at the research of fruit and vegetable image recognition,proposing an improved Alex Net network structure and an embedded fruit and vegetable image recognition App is designed based on Android system platform.This paper selects Alex Net as the core network framework,analyzes the structure optimization modules of Inception and Res Net,introduces different scales convolution kernel,and proposes an improved Alex Net network structure with multi-channel convolution and "shortcut" structure to realize the recognition function,the results show that the rate of the improved Alex Net is 96.99%,and the recognition time is2.45 s.Build the Android system development environment and configure the Open CV function library to realize the call of the fruit and vegetable recognition algorithm,build the fruit and vegetable identification information database through the SQLite database,call the camera function in the Android system to complete the design of the fruit and vegetable image recognition App,the results show that the rate of embedded system is 94.00% and the recognition time is 3.39 s.In this paper,it tests the improved Alex Net network model in several comparative experiments and compare the results of test between Embedded System and PC.The experiment results show that the network structure proposed in this paper can be better applied to embedded system equipment and realize the recognition function of embedded system of fruits and vegetables.