Analysis of Image Recognition by Multiple-Valued Neural Network for Automatic Strawberry Sorting System

摘要

This paper discusses a new technique of image recognition using a multiple-valued neural network (MVNN), for automatic strawberry sorting system. As we know that MVNN directly recognizes a raw image is difficult because of an image of strawberry fruit consisting of many multiple-valued picture elements. Here, firstly consider a method to extract image features from the strawberry fruit image and introduce the MVNN as well as its algorithm. It also describes that if a MVNN can be trained by a set of learning pattern successfully, the learnt MVNN would have the ability of recognizeing strawberry fruit image. Moreover, this paper also constructed a software of judging size and shape by learnt MVNN for the automatic strawberry sorting system.