X射线透射煤矸智能识别方法

王文鑫; 黄杰; 王秀宇; 史玉林; 吴高昌

doi:10.13272/j.issn.1671-251x.18037

X射线透射煤矸智能识别方法

doi: 10.13272/j.issn.1671-251x.18037

王文鑫^1,,
黄杰¹,
王秀宇²,
史玉林²,
吴高昌^{1, 3, ,}

1.
东北大学流程工业综合自动化国家重点实验室, 辽宁沈阳　110819
2.
沈阳隆基电磁科技股份有限公司, 辽宁沈抚新区　113122
3.
东北大学工业人工智能研究院, 辽宁沈阳　110819

基金项目: 国家自然科学基金青年科学基金项目 (62103092)；教育部中央高校基本科研业务费优秀青年科技人才培育项目 (N2108001) 。

详细信息

作者简介:
王文鑫（1997—），女，黑龙江齐齐哈尔人，硕士研究生，研究方向为图像处理与计算机视觉，E-mail：2102041@stu.neu.edu.cn

通讯作者:
吴高昌（1991—），男，安徽淮南人，副教授，博士，研究方向为图像处理与计算机视觉、光场成像与处理、异常工况智能预测、深度学习，E-mail：wugc@mail.neu.edu.cn

中图分类号: TD67
计量
- 文章访问数: 286
- HTML全文浏览量: 36
- PDF下载量: 65
- 被引次数: 0
出版历程
- 收稿日期: 2022-09-01
- 修回日期: 2022-11-09
- 网络出版日期: 2022-11-17

X-ray transmission intelligent coal-gangue recognition method

1.
State Key Laboratory of Synthetical Automation for Process Industries, Northeastern University, Shenyang 110819, China
2.
Longi Magnet Co., Ltd., Shenfu Reform and Innovation Demonstration Zone 113122, China
3.
Institute of Industrial Artificial Intelligence, Northeastern University, Shenyang 110819, China

摘要

摘要: 煤矸图像识别是基于伪双能X射线透射（XRT）的煤矸分选技术重要环节。受煤矸紧贴或遮挡导致煤矸图像难以分割和基于人工阈值判别易导致煤矸分类识别错误影响，现有的煤矸识别方法精度不高。提出一种XRT煤矸智能识别方法。采用感受野模块（RFB）与U−Net模型相结合的模型（RFB+U−Net模型）实现伪双能X射线煤矸图像有效分割，解决了因煤矸紧贴或遮挡情况而影响识别精度的问题；以煤矸图像灰度特征中的低能图像灰度最小值、纹理特征中的低能图像锐化最小值和锐化均差为煤矸识别特征，采用多层感知机（MLP）模型实现煤矸识别。实验表明：RFB+U−Net模型的煤矸分割准确率、煤矸粒度精度、煤矸像素均交并比等指标及图像分割效果均优于活动区域模型、U−Net模型、SegNet模型，且模型推理时间较短，满足煤矸图像分割实时性要求；MLP模型隐藏层数量为8时，在2组测试集下的煤矸识别平均准确率均为87%以上；在相同数据集及实验条件下，MLP模型的煤矸识别平均准确率、排矸率均高于基于贝叶斯分类器、支持向量机、逻辑回归、决策树、梯度提升决策树、K近邻算法的模型，且矸石带煤率不超过3%，满足实际煤矸干法分选要求。
- 煤矸分选 /
- 煤矸识别 /
- 煤矸图像分割 /
- X 射线透射 /
- 多层感知机 /
- 机器视觉
Abstract: The coal-gangue image recognition is an important part of coal-gangue separation technology based on pseudo dual energy X-ray transmission (XRT). However, it is difficult to segment the coal-gangue image due to the close proximity or occlusion of coal-gangue, and it is easy to cause classification and recognition errors of coal-gangue based on artificial threshold discrimination. Due to the above influence, existing coal-gangue recognition methods have low precision. In this paper, an X-ray transmission intelligent coal-gangue recognition method is proposed. A U-Net model combined with the receptive field block (RFB) is used to realize the effective segmentation of the pseudo dual energy X-ray coal-gangue image, which is termed as RFB + U-Net model. The problem that the recognition precision is affected by the close proximity or shielding of coal-gangue is solved. The recognition features of coal-gangue are the minimum gray value of the low-energy image in the gray level features of coal-gangue image, and the minimum value and the average difference of sharpened low-energy image in the texture features. A multi layer perceptron (MLP) model is used to realize coal-gangue recognition. Experimental results show that the RFB+U-Net model is superior to the active contour model, U-Net model and SegNet model in terms of coal-gangue segmentation accuracy, coal-gangue particle size precision, coal-gangue pixel mean intersection ratio and image segmentation effect. The reasoning time of the model is short, meeting the real-time requirements of coal-gangue image segmentation. When the number of hidden layers in the MLP model is 8, the average coal-gangue recognition accuracy under two test sets is more than 87%. Under the same data set and experimental conditions, the average recognition accuracy and gangue removal rate of the MLP model are higher than those based on Bayesian classifier, support vector machine, logic regression, decision tree, gradient boosting decision tree and K-nearest neighbor algorithm. The coal carrying rate of gangue shall not exceed 3%, meeting the requirements of actual dry coal-gangue separation.
- coal-gangue separation /
- coal-gangue recognition /
- coal-gangue image segmentation /
- X-ray transmission /
- multilayer perceptron /
- machine vision

HTML全文

图 1 RFB+U−Net模型结构

Figure 1. Releptive field block (RFB)+U-Net model structure

下载: 全尺寸图片幻灯片

图 2 MLP模型结构

Figure 2. Multilayer perceptron(MLP) model structures

下载: 全尺寸图片幻灯片

图 3 不同图像分割模型的煤矸图像分割结果

Figure 3. Coal-gangue image segmentation results of different image segmentation models

下载: 全尺寸图片幻灯片

图 4 MLP模型隐藏层数量消融实验结果

Figure 4. Ablation experiment results of hidden layer number of MLP models

下载: 全尺寸图片幻灯片

图 5 不同隐藏层数量的MLP模型收敛性能对比

Figure 5. Convergence performance comparison of MLP models with different hidden layer number

下载: 全尺寸图片幻灯片

图 6 不同煤矸识别模型的评价结果

Figure 6. Evaluation results of different coal-gaugue recognition models

下载: 全尺寸图片幻灯片

表 1 不同图像分割模型评价指标对比

Table 1. Comparison of evaluation indexes of different image segmentation models

模型	准确率/%	粒度精度/%	均交并比/%	推理时间/s
ACM	78.86	95.18	96.40	8.956 0
U−Net	95.65	94.46	95.92	0.045 3
SegNet	94.01	94.43	94.85	0.181 0
同等感受野U−Net	95.11	93.07	96.29	0.048 9
RFB+U−Net	96.31	95.65	96.62	0.047 2

下载: 导出CSV

参考文献(15)

[1]	LI C, KAO C Y, GORE J C, et al. Implicit active contours driven by local binary fitting energy[C]. IEEE Conference on Computer Vision and Pattern Recognition, Minneapolis, 2007: 1-7.
[2]	SHELHAMER E, LONG J, DARRELL T. Fully convolutional networks for semantic segmentation[C]. IEEE Conference on Computer Vision and Pattern Recognition, Boston, 2015: 3431-3440.
[3]	BADRINARAYANAN V,KENDALL A,CIPOLLA R,et al. SegNet:a deep convolutional encoder-decoder architecture for image segmentation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,2017,39(12):2481-2495. doi: 10.1109/TPAMI.2016.2644615
[4]	RONNEBERGER O, FISCHER P, BROX T. U-net: convolutional networks for biomedical image segmentation[C]. International Conference on Medical Image Computing and Computer-Assisted Intervention, Munich, 2015: 234-241.
[5]	KETELHODT L V,BERGMANN C. Dual energy X-ray transmission sorting of coal[J]. Journal of the South African Institute of Mining & Metallurgy,2010,110(7):371-378.
[6]	郭永存, 王鑫泉, 王爽, 等. 一种基于交叉算法边缘检测理论和视觉特征的煤矸识别方法: CN114066861A[P]. 2022-02-18. GUO Yongcun, WANG Xinquan, WANG Shuang, et al. A coal-gangue recognition method based on edge detection theory and visual characteristics of crossover algorithm: CN114066861A[P]. 2022-02-18.
[7]	李太友, 冯化一, 陈曦, 等. 目标检测网络训练方法及基于目标检测网络煤矸识别方法: CN112990350A[P]. 2021-04-12. LI Taiyou, FENG Huayi, CHEN Xi, et al. Training method of target detection network and gangue identification method based on target detection network: CN112990350A[P]. 2021-04-12.
[8]	王闰泽,郎利影,席思星. 用于智能煤矸分选机器人的改进型VGG网络煤矸识别模型[J]. 煤炭技术,2022,41(1):237-241. doi: 10.13301/j.cnki.ct.2022.01.059 WANG Runze,LANG Liying,XI Sixing. Improved VGG network coal gangue recognition model for intelligent coal gangue sorting robot[J]. Coal Technology,2022,41(1):237-241. doi: 10.13301/j.cnki.ct.2022.01.059
[9]	GARDNER M W,DORLING S R. Artificial neural networks (the multilayer perceptron)-a review of applications in the atmospheric sciences[J]. Atmospheric Environment,1998,32(14/15):2627-2636.
[10]	LETHIKIM N,NGUYENTRANG T,VOVAN T. A new image classification method using interval texture feature and improved Bayesian classifier[J]. Multimedia Tools and Applications,2022,81(25):36473-36488. doi: 10.1007/s11042-022-13531-6
[11]	ABDULLAH D M,ABDULAZEEZ A M. Machine learning applications based on SVM classification:a review[J]. Qubahan Academic Journal,2021,1(2):81-90. doi: 10.48161/qaj.v1n2a50
[12]	DREISEITL S,OHNO-MACHADO L. Logistic regression and artificial neural network classification models:a methodology review[J]. Journal of Biomedical Informatics,2002,35(5/6):352-359.
[13]	RANGANATHAN S, NAKAI K, SCHONVACH C. Encyclopedia of bioinformatics and computational biology: ABC of bioinformatics[M]. Elsevier, 2018.
[14]	SUN Rui,WANG Guanyu,ZHANG Wenyu,et al. A gradient boosting decision tree based GPS signal reception classification algorithm[J]. Applied Soft Computing,2020,86:105942. doi: 10.1016/j.asoc.2019.105942
[15]	RAJANI K L V, PADMA S Y. Classification of arrhythmia beats using optimized K-nearest neighbor classifier[M]//UDGATA S K, SETHI S, SRIRAMA S N. Intelligent systems. Singapore: Springer, 2021.