深度学习在选矿领域的应用进展  

Application of deep learning in mineral processing field

作　　者：田妞 黄宋魏[1] 和丽芳[1] 杨社平 字佳林 黄斌 Tian Niu;Huang Songwei;He Lifang;Yang Sheping;Zi Jialin;Huang Bin(Faculty of Land and Resources Engineering,Kunming University of Science and Technology,Kunming Yunnan 650093,China;Key Laboratory of Sanjiang Metallogenesis and Resource Exploration and Utilization,MNR,Kunming Yunnan 650200,China;Yunnan Provincial Bureau of Geology and Mineral Exploration and Development Center Laboratory,Kunming Yunnan 650200,China;Yunnan Phosphate Chemical Group Co.,Ltd.,National Engineering Research Center of Phosphate Resources Development and Utilization,Kunming Yunnan 650600,China)

机构地区：[1]昆明理工大学国土资源工程学院,云南昆明650093 [2]自然资源部三江成矿作用及资源勘查利用重点实验室,云南昆明650200 [3]云南省地质矿产勘查开发局中心实验室,云南昆明650200 [4]云南磷化集团有限公司,国家磷资源开发利用工程技术研究中心,云南昆明650600

出　　处：《化工矿物与加工》2023年第8期69-74,82,共7页Industrial Minerals & Processing

基　　金：云南省科技厅基金项目(202101AT070277);云南省万人计划人才项目(CG22166F219A);国家磷资源开发利用工程技术研究中心开放基金项目(NECP2022-11)。

摘　　要：深度学习作为一门新兴技术,可以很好地处理非线性问题以及排除由主观经验导致的稳定性低等问题,已成为选矿领域的研究热点。阐述了基于计算机图像处理技术的深度学习原理,综述了深度学习在矿物识别、碎矿和磨矿、浮选、重选、磁选等方面的研究进展,总结了其优势和不足,并对未来的研究方向提出了若干建议。As an emerging technology,deep learning is likely to effectively handle nonlinear issues and eliminate low stability caused by subjective experience and has become a focus of research in mineral processing field.In this article the principles of deep learning based on computer image processing technology was presented.The research progress and application of deep learning in mineral identification,ore crushing and grinding,flotation,gravity separation,magnetic separation and other processes were reviewed,in which both advantages and disadvantages of it were summarized and some recommendations on further study were made.

关 键 词：选矿 深度学习 图像处理 智能化 矿物识别 磨矿参数优化 

分 类 号：TD97[矿业工程—选矿]