计算机与现代化 ›› 2020, Vol. 0 ›› Issue (10): 58-63.

• 图像处理 • 上一篇    下一篇

一种高精度微孔板浑浊度识别网络

  

  1. (1.华南农业大学数学与信息学院,广东广州510642;2.华南农业大学兽医学院,广东广州510642)
  • 出版日期:2020-10-14 发布日期:2020-10-14
  • 作者简介:李西明(1974—),男,山东临清人,副教授,博士,研究方向:图像处理,信息安全,E-mail: liximing@scau.edu.cn; 马李晓(1997—),男,本科生,研究方向:图像处理,E-mail: 1769792909@qq.com; 曾晓银(1995—),女,硕士研究生,研究方向:图像识别,E-mail: 1649028910@qq.com; 王璇(1997—),男,硕士研究生,研究方向:网络安全密码学,E-mail: 1422916521@qq.com; 孙坚(1984—),男,教授,博士,研究方向:细菌耐药及进化机制,耐药性逆转,E-mail: jiansun@scau.edu.cn; 郭玉彬(1973—),女,山东临清人,副教授,博士,研究方向:图像处理,信息安全,E-mail: guoyubin@scau.edu.cn。
  • 基金资助:
    国家重点研发计划项目(2016YFD0501300); 国家基金海外合作重点项目(30520103918); 广东省农业厅省级乡村振兴战略专项项目(粤农计[2018]54号)

A High Precision Microporous Plate Turbidity Identification Network

  1. (1. College of Mathematics and Informatics, South China Agricultural University, Guangzhou 510642, China;
    2. College of Veterinary Medicine, South China Agricultural University, Guangzhou 510642, China)
  • Online:2020-10-14 Published:2020-10-14

摘要: 提出一种基于卷积神经网络的高精度微孔板浑浊度分类算法。该算法主要将传统图像处理技术与卷积神经网络技术相结合,通过传统图像处理算法将圆孔从自然拍摄的微孔板图像中切割下来,并将切割下来的圆孔图像制作成圆孔数据集,用于网络模型的训练、评估和测试。同时,通过深度学习技术,设计并训练多个基于深度可分离卷积核的卷积神经网络模型,然后筛选出评估准确率最高的浑浊度分类模型,应用于圆孔识别系统,从而可提高研究人员的工作效率。

关键词: 图像分类, 深度学习, 卷积神经网络

Abstract: A high precision microporous plate turbidity classification algorithm based on convolutional neural network is proposed. This algorithm mainly combines the traditional image processing technology with the convolutional neural network technology. Through the traditional image processing algorithm, round holes are cut from the microporous plate images taken naturally, and the cut round hole images are made into round hole data sets for the training, evaluation and testing of network models. At the same time, through the deep learning technology, multiple convolutional neural network models based on the depth-separable convolution kernel are designed and trained. Then, the turbidity classification model with the highest evaluation accuracy is selected and applied to the circular hole identification system, thus improving the work efficiency of researchers.

Key words: image classification, deep learning, convolutional neural network