基于CMI-UNet的煤泥浮选泡沫分割算法

Coal flotation froth segmentation algorithm based on CMI-UNet

摘要

浮选泡沫边缘分割与识别是选煤厂核心工艺过程参数预测与优化的关键技术, 对于煤泥浮选智能化建设具有重要意义。目前, 基于U型结构网络(U-Net)的分割算法存在高层语义特征与底层信息的语义鸿沟和简单拼接带来的歧义等问题。本研究基于通道多尺度集成U型网络(CMI-UNet)提出一种新的煤泥浮选泡沫分割算法, 该算法在U-Net网络基础上, 增加了通道交叉融合Transformer模块、多分辨率融合模块和信息瓶颈模块, 从而有效弥合潜在的语义歧义, 提升特征表达能力、增强算法对复杂多尺度信息的捕捉和表达能力。在自建数据集上的实验结果表明, 所提CMI-UNet算法对煤泥浮选泡沫具有更好的分割效果。

Abstract

Flotation froth edge segmentation is a key technology for the prediction and optimization of the core process parameters in coal preparation plants. It is significant to the intelligent construction of the coal flotation. So far, the segmentation algorithms based on U-Net still have the following problems: the semantic gap existing between the high-level semantic features and underlying information features, and semantic ambiguity due to simple splicing. This study proposed a new coal flotation froth segmentation algorithm based on CMI-UNet. By incorporating the channel cross fusion transformer (CCFT) module, multi-resolution fusion (MRF) module and information bottleneck (IB) module based on U-Net, this network effectively bridges the potential semantic ambiguities, enhances the ability of feature expression, and strengthens its capacity to capture and express complex multiscale information. The experimental results on self-built dataset show that the proposed CMI-UNet segmentation algorithm has a better segmentation effect on coal flotation froth.

关键词

煤泥浮选 / 泡沫图像 / 深度学习 / 语义分割 / U型结构网络

Key words

coal flotation / froth image / deep learning / semantic segmentation / U-Net

参考文献

[1] ALDRICH C , AVELAR E , LIU X . Recent advances in flotation froth image analysis[J]. Minerals Engineering, 2022, 188: 107823-107839.

[2] JAHEDSARAVANI A , MASSINAEI M , ZARIE M . Measurement of bubble size and froth velocity using convolutional neural networks[J]. Minerals Engineering, 2023, 204: 108400-108414.

[3] WANG Z H , WANG Z Z . A generic approach for cell segmentation based on gabor filtering and area constrained ultimate erosion[J/OL]. Artificial Intelligence in Medicine, 2020, 107.DOI: 10.1016/j.artmed.2020.101929.

[4] ZHANG H , TANG Z H , XIE Y F , et al. A watershed segmentation algorithm based on an optimal marker for bubble size measurement[J]. Measurement, 2019, 138: 182-193.

[5] PENG C , LIU Y K , GUI W H , et al. Bubble image segmentation based on a novel watershed algorithm with an optimized mark and edge constraint[J/OL]. IEEE Transactions on Instrumentation and Measurement, 2021, 71.DOI: 10.1109/TIM.2021.3129873.

[6] LIANG X M , TIAN T , LIU W T , et al. Flotation froth image segmentation based on highlight correction and parameter adaptation[J]. Mining, Metallurgy and Exploration, 2020, 37(2): 467-474.

[7] MINAEE S , BOYKOV Y , PORIKLI F , et al. Image segmentation using deep learning: a survey[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2021, 44(7): 3523-3542.

[8] RONNEBERGER O , FISCHER P , BROX T . U—net: Convolutional networks for biomedical image segmentation[C]// Medical Image Computing and Computer Assisted Intervention. 2015: 234-241.

[9] JU Y Y, WU L C, LI M, et al. A novel hybrid model for flow image segmentation and bubble pattern extraction[J/OL]. Measurement, 2022, 192.DOI: 10.1016/j.measurement.2022.110861.

[10] WANG S J , GUAN H F , WANG Y Q , et al. Intelligent recognition of gas liquid two phase flow based on optical image[J]. The International Arab Journal of Information Technology, 2023, 20(4): 609-617.

[11] JHA D , SMEDSRUD P H , RIEGLER M A , et al. Resunet++: An advanced architecture for medical image segmentation[C]// 2019 IEEE International Symposium on Multimedia. 2019: 225-230.

[12] 代茵, 刘维宾, 董昕阳, 等 . 基于注意力机制的U—Net脑脊液细胞分割[J]. 东北大学学报(自然科学版), 2022, 43(7): 944-950.

[13] DAI Yin , LIU Weibin , DONG Xinyang , et al. U—Net CSF cells segmentation based on attention mechanism[J]. Journal of Northeastern University(Natural Science), 2022, 43(7): 944-950.

[14] 马思珂, 赵萌, 石凡, 等 . 注意力驱动的细胞团簇细胞核分割算法[J]. 西安电子科技大学学报, 2022, 49(2): 198-206.

[15] MA Sike , ZHAO Meng , SHI Fan , et al. Attention driven nuclei segmentation method for cell clusters[J]. Journal of Xidian University, 2022, 49(2): 198-206.

[16] 唐朝晖, 郭俊岑, 张虎, 等 . 基于改进I—Attention U—Net的锌浮选泡沫图像分割算法[J]. 湖南大学学报(自然科学版), 2023, 50(2): 12-22.

[17] TANG Zhaohui , GUO Juncen , ZHANG Hu , et al. Zinc flotation froth image segmentation algorithm based on improved I—Attention U—Net[J]. Journal of Hunan University(Natural Science Edition), 2023, 50(2): 12-22.

[18] 徐宏祥, 李神舟, 徐培培 . 基于DA—Attention U—Net编码—解码结构的浮选矿浆相气泡图像分割[J]. 有色金属(选矿部分), 2024(6): 106-115.

[19] XU Hongxiang , LI Shenzhou , XU Peipei . Segmentation of flotation pulp phase bubbles image based on DA—Attention U—Net[J]. Nonferrous Metals(Mineral Processing Section), 2024(6): 106-115.

[20] WANG H , CAO P , WANG J , et al. UCTransNet: Rethinking the skip connections in U—Net from a channel wise perspective with transformer[C]// Proceedings of the AAAI Conference on Artificial Intelligence. 2022, 36(3): 2441-2449.

[21] MA M J , XIA H Y , TAN Y M , et al. HT—Net: Hierarchical context attention transformer network for medical CT image segmentation[J]. Applied Intelligence, 2022, 52(9): 10692-10705.

[22] HUANG S Q , LI J N , XIAO Y Z , et al. RTNet: Relation transformer network for diabetic retinopathy multilesion segmentation[J]. IEEE Transactions on Medical Imaging, 2022, 41(6): 1596-1607.

[23] JIANG S , LI J L , HUA Z . DPCFN: Dual path cross fusion network for medical image segmentation[J/OL]. Engineering Applications of Artificial Intelligence, 2022, 116.DOI: 10.1016/j.engappai.2022.105420.

[24] WANG C , WANG L , WANG N Q , et al. CFATransUnet: Channel wise cross fusion attention and transformer for 2D medical image segmentation[J/OL]. Computers in Biology and Medicine, 2024, 168.DOI: 10.1016/j.compbiomed.2023.107803.

[25] DHEIVYA I , SARAVANA KUMAR G . Multi resolution neural network for road scene segmentation[C]// International Conference on Deep Learning, Artificial Intelligence and Robotics.Cham: Springer Nature Switzerland, 2023: 195-205.

[26] HU Y T , HUANG X , LUO X Y , et al. Learning foreground information bottleneck for few shot semantic segmentation[J/OL]. Pattern Recognition, 2024, 146.DOI: 10.1016/j.patcog.2023.109993.

基金资助

山东省自然科学基金项目(ZR2023MF112)