作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2020, Vol. 46 ›› Issue (9): 221-225. doi: 10.19678/j.issn.1000-3428.0056094

• 图形图像处理 • 上一篇    下一篇

基于改进残差网络的果实病害分类方法

肖经纬, 田军委, 王沁, 程希希, 王佳   

  1. 西安工业大学 机电工程学院, 西安 710021
  • 收稿日期:2019-09-23 修回日期:2019-10-29 发布日期:2019-10-31
  • 作者简介:肖经纬(1995-),男,硕士研究生,主研方向为人工智能、图像识别;田军委,教授、博士;王沁、程希希、王佳,硕士研究生。
  • 基金资助:
    陕西省科技计划项目(2020NY-148);西安市未央科技计划项目(201840,201920)。

Fruit Disease Classification Method Based on Improved Residual Network

XIAO Jingwei, TIAN Junwei, WANG Qin, CHENG Xixi, WANG Jia   

  1. School of Mechanical and Electrical Engineering, Xi'an Technological University, Xi'an 710021, China
  • Received:2019-09-23 Revised:2019-10-29 Published:2019-10-31

摘要: 传统的残差网络在果实病害分类中存在层数较多,以及在实际应用中有参数冗余的问题,且原始损失函数对具有相似特征的病害容易造成错误识别。为解决果害分类中参数过多及相似样本区分度低的问题,提出一种改进的残差网络结构,以降低残差块数量与卷积核数量来减少卷积层参数。同时,在原始损失函数中加入类间相似惩罚项来扩大不同类间距,以提高对病害的分类准确率。实验结果表明,相比原始的残差网络,改进后的残差网络降低约25%的参数量,改进后损失函数的识别准确率达到92.76%。

关键词: 深度学习, 残差网络, 图像分类, 果实病害, 损失函数

Abstract: The number of layers and parameters of traditional residual networks are redundant in the practical application of fruit disease classification,and the original loss function is easy to misidentify similar disease.In order to solve the problem of redundant parameters and low discrimination of similar samples in fruit classification,this paper proposes an improved residual network structure to reduce the number of residual blocks and convolution kernels to reduce the parameters of the convolution layer.At the same time,the inter-class similarity penalty term is added into the original loss function to widen the distance between different classes,so as to improve the classification accuracy of diseases.Experimental results show that compared with the original residual network,the improved residual network reduces the amount of parameters by about 25%,and the recognition accuracy of the improved loss function reaches 92.76%.

Key words: deep learning, residual network, image classification, fruit disease, loss function

中图分类号: