摘 要:偏光显微镜下的岩石薄片识别,不仅在分析和鉴定岩石结构和矿物成分中具有至关重要的作用,而且在矿产资源勘探(尤其是固体金属矿产资源勘探)中同样发挥着不可估量的作用。目前,显微镜下对岩石薄片的观察主要依靠专业人员的肉眼来识别岩石薄片中的矿物成分,费时耗力。为此,针对偏光显微镜下岩石数据样本的特征,构建一种基于 ResNeXt 技术的岩石薄片识别与分类模型,借助该模型分析岩石薄片可大大节约人力成本和时间成本。
关键词:岩石薄片;ResNeXt;图像分类
DOI:10.19850/j.cnki.2096-4706.2022.24.016
中图分类号:TP18 文献标识码:A 文章编号:2096-4706(2022)24-0066-03
Rock Slice Recognition under Microscope Based on Deep Learning
ZHU Jun, ZHANG Junkun, CHEN Yao, LI Yanlin
(School of Mathematics and Computers (Big Data Science), Panzhihua University, Panjihua 617000, China)
Abstract: The identification of rock slices under polarizing microscope not only plays an important role in analyzing and identifying rock structure and mineral composition, but also plays an inestimable role in the exploration of mineral resources (especially the exploration of solid metal mineral resources). At present, the observation of rock slices under microscope mainly depends on professionals' naked eye to identify the mineral composition in rock slices, which is time-consuming and labor-intensive. Therefore, according to the characteristics of rock data samples under polarizing microscope, a rock slice recognition and classification model based on ResNeXt technology is constructed. The analysis of rock slices with this model can greatly save labor costs and time costs.
Keywords: rock slice; ResNeXt; image classification
参考文献:
[1] 胡祺 . 融合多维信息的岩石薄片图像深度学习分类方法[D]. 杭州:浙江大学,2019.
[2] 周永章,左仁广,刘刚,等.数学地球科学跨越发展的十年:大数据、人工智能算法正在改变地质学 [J]. 矿物岩石地球化学通报,2021,40(3):556-573+777.
[3] 程国建,郭文惠,范鹏召 . 基于卷积神经网络的岩石图像分类 [J]. 西安石油大学学报:自然科学版,2017,32(4):116-122.
[4] 张野,李明超,韩帅 . 基于岩石图像深度学习的岩性自动识别与分类方法 [J]. 岩石学报,2018,34(2):333-342.
[5] FU J Z. Application of Modified Inception-ResNet and CondenseNetinLung Nodule Classification [C]// 3rd International Conference on Computer Engineering, Information Science & Application Technology(ICCIA 2019). 重庆:[ 出版者不详 ].2019:197-205.
[6] 冯雅兴,龚希,徐永洋,等 . 基于岩石新鲜面图像与孪生卷积神经网络的岩性识别方法研究 [J]. 地理与地理信息科学,2019,35(5):89-94.
[7] HE K M,ZHANG X Y,REN S Q,et al. Deep residual learning for image recognition [C]//2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).Las Vegas:IEEE,2016:770-778.
[8] SZEGEDY C,LIU W,JIA Y Q,et al. Going deeper with convolutions [C]//2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).Boston:[s.n.]. 2015:1-9.
[9] JOBSON D J,RAHMAN Z. A multiscale retinex for bridging the gap between color images and the human observation of scenes [J]. IEEE transactions on image processing,1997(7):965-976.
作者简介:朱君(2001.08—),男,汉族,四川广安人,本科在读,研究方向:机器视觉;李炎林(1999.03—),男,汉族,四川广元人,本科在读,研究方向:机器视觉;陈尧(1983.10—),男,汉族,四川西昌人,讲师,博士,研究方向:人工智能;张俊坤(1980.10—),男,汉族,四川攀枝花人,讲师,硕士,研究方向:数据挖掘。