光子学报 ›› 2020, Vol. 49 ›› Issue (10): 1015001-1015001.doi: 10.3788/gzxb20204910.1015001

• 计算机视觉 • 上一篇    下一篇

基于双模态卷积神经网络自适应迁移学习的浮选工况识别

廖一鹏1(), 杨洁洁1, 王志刚2, 王卫星1()   

  1. 1.福州大学 物理与信息工程学院,福州 350108
    2.福建金东矿业股份有限公司,福建 三明 365101
  • 收稿日期:2020-04-09 接受日期:2020-06-17 出版日期:2020-10-25 发布日期:2020-10-13
  • 通讯作者: 王卫星 E-mail:fzu_lyp@163.com;wxwwx@fzu.edu.com
  • 作者简介:廖一鹏(1982—),男,讲师,博士研究生,主要研究方向为机器视觉技术、图像处理与模式识别. Email: fzu_lyp@163.com
  • 基金资助:
    国家自然科学基金(61471124);福建省自然科学基金(2019J01224);福建省中青年教师教育科研项目(JT180056)

Flotation Performance Recognition Based on Dual-modality Convolutional Neural Network Adaptive Transfer Learning

Yi-peng LIAO1(), Jie-jie YANG1, Zhi-gang WANG2, Wei-xing WANG1()   

  1. 1.College of Physics and Information Engineering,Fuzhou University,Fuzhou 350108,China
    2.Fujian Jindong Mining Co. Ltd. ,Sanming,Fujian 365101,China
  • Received:2020-04-09 Accepted:2020-06-17 Online:2020-10-25 Published:2020-10-13
  • Contact: Wei-xing WANG E-mail:fzu_lyp@163.com;wxwwx@fzu.edu.com
  • Supported by:
    National Natural Science Foundation of China(61471124);Fujian Provincial Natural Science Foundation(2019J01224);Fujian Provincial Education Scientific Research Project for Middle-aged and Young Teachers(JT180056)

摘要:

为提高小规模训练集下CNN特征驱动的浮选工况识别效果,提出一种基于泡沫红外与可见光图像CNN特征提取及自适应迁移学习的工况识别方法.首先构建基于AlexNet的双模态CNN特征提取及识别模型,并通过RGB-D大规模数据集对模型的结构参数进行预训练;其次,用多个串联的双隐层自编码极限学习机代替预训练模型的全连接层,实现对双模态CNN特征的融合及逐层抽象提取,然后通过核极限学习机映射到更高维空间进行决策;最后构建浮选小规模数据集对迁移后的模型进行训练,并改进量子狼群算法用于模型参数优化.实验结果表明:自适应迁移学习能够明显提高小样本数据集下的识别准确度,采用双模态CNN迁移学习较单模态CNN迁移学习的工况识别精度提高了3.06%,各工况的平均识别准确率达到96.83%,识别精度和稳定性较现有方法有较大提升.

关键词: 机器视觉, 浮选工况识别, 红外与可见光图像, 卷积神经网络, 迁移学习, 双隐层自编码极限学习机, 量子狼群算法

Abstract:

In order to improve the effect of CNN feature driven flotation performance recognition under small-scale training set, a method of flotation performance recognition based on adaptive transfer learning and CNN features extraction of foam infrared and visible images is proposed. Firstly, a dual-modality CNN feature extraction and recognition model based on AlexNet was constructed, and the structural parameters of the model were pre-trained through RGB-D large-scale data set. Secondly, a series of double hidden layer automatic encoder extreme learning machine is used to replace the full connection layer of the pre-training model, so that the dual-modality CNN features can be fused and abstracted layer by layer, and then the decision is made by mapping to higher dimensional space through the kernel extreme learning machine. Finally, the floatation small-scale data set is constructed to train the migrated model, and the improved quantum wolf pack algorithm is used for model parameter optimization. Experimental results show that, adaptive transfer learning can significantly improve the accuracy of recognition in small sample data sets, the accuracy of performance recognition using dual-modality CNN transfer learning is 3.06% higher than that of single-mode CNN transfer learning, and the average recognition accuracy of each working condition reached 96.86%. The accuracy and stability of flotation performance recognition is greatly improved compared with the existing methods.

Key words: Machine vision, Flotation performance recognition, Infrared and visible images, Convolutional neural network, Transfer learning, Double hidden layer automatic encoder extreme learning machine, Quantum wolf pack algorithm

中图分类号: