基于图像识别的赤铁矿反浮选泡沫颜色特征提取与分析

现代矿业 ›› 2025, Vol. 41 ›› Issue (09): 13-19.

基于图像识别的赤铁矿反浮选泡沫颜色特征提取与分析

代泽明马艺闻柳小波张艳兵李明娇王睿

辽宁科技大学矿业工程学院（智慧矿山研究院）

出版日期:2025-09-25 发布日期:2025-11-03

Color Feature Extraction and Analysis of Hematite Reverse Flotation Froth Based on Image Recognition

School of Mining Engineering（Intelligent Mine Research Institute）， University of Science and Technology Liaoning

Online:2025-09-25 Published:2025-11-03

摘要/Abstract

摘要： 为了实现对赤铁矿反浮选泡沫颜色特性的有效量化分析，并建立与浮选品位之间的关联模型，进行了基于多颜色空间协同分析与机器学习的泡沫颜色特征分析研究。通过分析赤铁矿反浮选泡沫颜色特点，提取了包括颜色矩、标准差和相对红值在内的65类初始颜色特征，并使用相关系数分析法从中筛选出与浮选品位显著相关的关键颜色特征，包括RGB颜色空间中G通道均值、YUV 颜色空间中 Y通道均值与 V通道偏度，以及灰度空间下的相对红值等。在此基础上，采用随机森林算法构建浮选品位预测模型并进行了验证试验，得到了均方误差0.47、决定系数0.73、平均绝对误差 0.51的结果；所提取的关键颜色特征能够有效表征赤铁矿反浮选泡沫颜色，且与浮选品位之间存在较强的相关性，可作为赤铁矿反浮选工况视觉辨识的有效工具，对开发基于泡沫颜色变化的赤铁矿浮选工况监控系统，提升浮选自动化与精细化控制水平具有重要意义。

关键词: 赤铁矿反浮选, 图像识别, 颜色特征, 泡沫状态量化

Abstract: In order to realize the effective quantitative analysis of the color characteristics of hematite reverse flotation foam and establish the correlation model with flotation grade，the analysis of foam color characteristics based on multi-color space collaborative analysis and machine learning is carried out. By an⁃ alyzing the color characteristics of hematite reverse flotation froth，65 kinds of initial color features includ⁃ ing color moment，standard deviation and relative red value are extracted，and the key color features signifi⁃ cantly related to flotation grade are selected by correlation coefficient analysis method，including the mean value of G channel in RGB color space，the mean value of Y channel and the skewness of V channel in YUV color space，and the relative red value in gray space. On this basis，the random forest algorithm is used to construct the flotation grade prediction model and the verification test is carried out. The results of mean square error of 0.47，determination coefficient of 0.73 and average absolute error of 0.51 are obtained.The key color features extracted can effectively characterize the color of hematite reverse flotation froth，and there is a strong correlation between the color and flotation grade.It can be used as an effective tool for visual identification of hematite reverse flotation conditions，which is of great significance for the development of hematite flotation condition monitoring system based on foam color change and the improvement of flotation automation and fine control level.

Key words: hematite reverse flotation, image recognition, color characteristics, foam state quantifica? tion

代泽明, 马艺闻, 柳小波, 张艳兵, 李明娇, 王睿. 基于图像识别的赤铁矿反浮选泡沫颜色特征提取与分析[J]. 现代矿业, 2025, 41(09): 13-19.

DAI Zeming MA Yiwen LIU Xiaobo ZHANG Yanbing LI Mingjiao WANG Rui. Color Feature Extraction and Analysis of Hematite Reverse Flotation Froth Based on Image Recognition[J]. Modern Mining, 2025, 41(09): 13-19.

扫码分享