浮选泡沫图像特征提取方法研究进展  

Research Progress of Flotation Froth Image Feature Extraction Method

作　　者：宛鹤[1] 陆笑科 屈娟萍 薛季玮 张崇辉[1] 王森[1] 卜显忠[1] WAN He;LU Xiaoke;QU Juanping;XUE Jiwei;ZHANG Chonghui;WANG Sen;BU Xianzhong(School of Resource Engineering,Xi′an University of Architecture and Technology,Xi′an 710055,Shaanxi,China;Oulu School of Mining,University of Oulu,Oulu FI-90014,Finland)

机构地区：[1]西安建筑科技大学资源工程学院,陕西西安710055 [2]奥卢大学奥卢矿业学院,芬兰奥卢FI-90014

出　　处：《中国钼业》2024年第1期1-8,共8页China Molybdenum Industry

基　　金：国家自然科学基金项目(编号:52274271,52074206,52104266)。

摘　　要：机器视觉作为设备操作人员的工具,在泡沫浮选设备的监测中得到了广泛的应用。利用泡沫图像数据集建立预测识别模型,以初级泡沫特征参数为输入,以品位和回收率等浮选指标为输出。根据是否需要手动提取浮选泡沫图像特征,可以将特征提取算法划分为两大类别:一种是基于颜色、形态特征等的传统手动特征提取方法,另一种是基于深度神经网络的自动特征提取方法。本文总结并归纳了近年来浮选泡沫图像特征提取算法领域的研究进展,分析了各种方法的优势和不足,对当前难以人工识别泡沫状态及实现浮选自动化提升浮选效率,具有一定的指导价值。As a tool for equipment operators,machine vision has been widely used in the monitoring of froth flotation equipment.A predictive identification model has been developed,utilizing a froth image dataset,with primary froth characteristic parameters as inputs and flotation indicators like grade and recovery as outputs.Depending on the necessity for manual extraction of flotation froth image features,the feature extraction algorithms can be divided into two main categories:one is traditional manual feature extraction methods that rely on aspects such as color and morphological features,and the other is automatic feature extraction methods grounded in deep neural networks.This paper summarizes the research progress in the area of flotation froth image feature extraction algorithms over recent years,while also critically examining the advantages and drawbacks of various methods it has certain guiding value for the curront difficulty in manually indentifying foam status and realizing flotation automation to improve flotation efficiency.

关 键 词：泡沫浮选 泡沫图像 机器视觉 泡沫图像特征 

分 类 号：TD923.7[矿业工程—选矿] TP391.41[自动化与计算机技术—计算机应用技术]