铁矿粉铁含量的高光谱分析和定量反演研究

Study on Hyperspectral Analysis of Iron Ore Powder and Quantitative Inversion of Iron Grade

高光谱检测铁矿粉铁含量是一种新技术,具有无损、高效的检测优势。为了研究高光谱识别技术对铁矿选矿粉铁含量测定的精确度,选取了铁矿选矿粉作为测试样本,采集了不同铁含量铁矿选矿粉的高光谱曲线,经过平滑去噪和光谱特征提取后,开展铁含量光谱拟合实验和铁含量的反演研究。研究结果表明:选矿粉铁含量与高光谱曲线高度相关性波段为Fe的强吸收位置517~520 nm和873~888 nm;在吸收位置520 nm左右,进行了基于最小二乘的选矿粉铁含量的建模反演,反演实验结果的拟合度为0.9885,相对误差为7.26%,说明利用高光谱技术进行铁矿粉铁含量检测准确度较高,为高光谱检测铁含量的实际应用提供了理论基础。

Hyperspectral detection of iron grade in iron ore dressing powder is a new technology,which has the advantages of nondestructive and efficient test.Some iron ore powders with different grades were selected as test samples in order to test the accuracy of identifying iron grade of iron ore dressing powder by hyperspectral spectroscopy.The hyperspectral curves of the samples were collected,after smoothing denoising and spectral feature extraction,the iron grade spectrum fitting test and iron grade inversion modeling were carried out.The results show that the high correlation between iron grade of iron ore powder and hyperspectral curve is the strong absorption position of Fe,about 517~520 nm and 873~888 nm;The modeling inversion of iron ore grade at about 520 nm based on the least square method is carried out.The determination coefficient of the inversion test results is 0.9885,and the relative error is 7.26%,it shows that the detection accuracy of iron grade of iron ore powder by hyperspectral technology is high,which provides a theoretical basis for the practical application of hyperspectral test of iron grade.