当前位置>主页 > 期刊在线 > 计算机技术 >

计算机技术22年14期

基于 RegNet 网络的岩石图像模式识别
林海涛,郑群浩,林嘉仪,陈沛逸
(韩山师范学院 数学与统计学院,广东 潮州 521041)

摘  要:建立了以岩石图片为输入的深度学习模型,以实现岩石类型的智能识别。该模型基于 RegNet 网络结构,通过探索网络设计空间特征,高效获取模型的最优参数组合。为了提高类型的识别率,在数据预处理中采用了置信学习和数据增强等方法和技巧,在训练阶段采用了迁移学习方法。实验表明,该模型的识别准确率达到 89.94%,在某些类别的召回率甚至高达 100%。


关键词:RegNet 网络;模式识别;置信学习;数据增强;迁移学习



DOI:10.19850/j.cnki.2096-4706.2022.014.015


基金项目:2021博士启动项目(QD202129);2021 年度教育厅教育科学规划项目(2021GXJK212)


中图分类号:TP18                                          文献标识码:A                                   文章编号:2096-4706(2022)14-0063-04


Rock Image Pattern Recognition Based on RegNet Network

LIN Haitao, ZHENG Qunhao, LIN Jiayi, CHEN Peiyi

(School of Mathematics and Statistics, Hanshan Normal University, Chaozhou 521041, China)

Abstract: This paper builds a deep learning model with inputting rock images to realize intelligent identification of rock types. The model is based on the RegNet network structure, efficiently obtains optimal parameter combination by exploring the network design spatial features. In order to improve the identification rate of types, this paper uses methods and techniques such as confidence learning and data augmentation in data preprocessing, and uses transfer learning method during the training phase. Experiments show that the identification accuracy of the model achieves 89.94%, and the recall rate for some types is even as high as 100%.

Keywords: RegNet network; pattern recognition; confidence learning; data augmentation; transfer learning


参考文献:

[1] SINGH N,SINGH T N,TIWARY A,et al. Textural identification of basaltic rock mass using image processing and neural network [J].Computational Geosciences,2010,14(2):301-310.

[2] MŁYNARCZUK M,GÓRSZCZYK A,ŚLIPEK B. The application of pattern recognition in the automatic classification of microscopic rock images [J].Computers & Geosciences,2013,60:126-133.

[3] LIU Y,GUO C,CHENG G J. Rock classification based on images color spaces and artificial neural network [C]//2014 Fifth International Conference on Intelligent Systems Design and Engineering Applications.Hunan:IEEE,2014:897-900.

[4] 李燕 . 基于深度学习的岩石图像识别 [D]. 北京:北京林业大学,2020.

[5] RADOSAVOVIC I,KOSARAJU R P,GIRSHICK R,et al. Designing network design spaces [C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).Seattle:IEEE, 2020:10425-10433.

[6] NORTHCUTT C G,JIANG L,CHUANG I L. Confident learning:Estimating uncertainty in dataset labels [J/OL].arXiv: 1911.00068 [stat.ML].[2022-05-17].https://arxiv.org/abs/1911.00068v2. 

[7] LEEVY J L,KHOSHGOFTAAR T M,BAUDER R A,etal. A survey on addressing high-class imbalance in big data [J].Journal of Big Data,2018,5(1):1-30.

[8] PAN S J L,YANG Q. A survey on transfer learning [J].IEEE Transactions on knowledge and data engineering,2009,22(10):1345-1359.


作者简介:林海涛(1982—),男,汉族,广东揭阳人,讲师,博士,主要研究方向:模式识别与数据挖掘;郑群浩(1998—),男,汉族,广东潮州人,本科在读,主要研究方向:大数据技术;林嘉仪(2000—),女,汉族,广东揭阳人,本科在读,主要研究方向:数据分析与数据挖掘;陈沛逸(2000—),男,汉族,广东广州人,本科在读,主要研究方向:数据分析与数据挖掘。