Open AccessRecognition of navel orange image with complex background based on residual network
Author(s) -
Xiaoxin Li,
Yumei Tan,
Xinxin Lu,
Bo Zhang
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1861/1/012050
Subject(s) - navel orange , residual , orange (colour) , navel , residual neural network , artificial intelligence , computer science , pattern recognition (psychology) , convolutional neural network , artificial neural network , algorithm , horticulture , biology , anatomy
In order to solve the problem that the detection effect of navel orange recognition in complex background is poor, we propose a navel orange recognition method based on residual network. In this study, the navel orange classification dataset was constructed and labeled, and the performance of five classic convolutional neural network models on this dataset was evaluated, including AlexNet, Improved LeNet, SqueezeNet, ResNet-18, GoogLeNet. The results present significant accuracy obtained by the ResNet-18 model, with accuracy of 98.27%, which is more suitable for navel orange image recognition in complex background.