摘要
网络度是衡量化纤丝线及化纤织物性能的重要指标之一,在生产车间中通常采用人工方式进行检测.为解决人工检测误检率较高的问题,提出一种基于语义信息增强的化纤丝线网络度并行检测方法.首先,为提升单根化纤丝线网络结点识别的准确度,使用基于MobileNetV2优化的主干网络结构提取语义信息,以提高模型的运算速度.在所提主干网络的基础上,设计语义信息增强模块和多级特征扩张模块处理主干网络的特征信息,同时,设计像素级注意力掩膜对特征信息进行加权和融合,以提高网络度检测的准确性.然后,为实现多根化纤丝线网络度的批量计算,基于所提语义信息增强算法,设计网络度并行检测方法.使用算法检测丝线网络结点,同时使用连通域分析及掩膜提取的方法并行检测,提取视野内每条丝线的独立区域.随后,将并行检测结果融合,以准确获取每根丝线的网络度检测结果.为验证所提方法的有效性,使用自主研发的网络度检测设备建立了化纤丝线数据集,并进行了实验验证.结果表明,所提出的方法能够有效地提高检测的准确性.
The interlacing degree serves as an important indicator for evaluating the performance of filament yarns and fabrics,typically detected manually in production workshop.To address the issues of high false detection rates in manual inspection,a parallel detection method for filament yarn interlacing degree based on semantic information enhancement is proposed.Firstly,to improve the recognition accuracy of interlacing nodes in a filament yarn,an improved backbone architecture based on MobileNetV2 is used for semantic information extraction to improve the computational speed of model.Building upon the proposed backbone architecture,semantic information enhancement module and multilevel feature dilated module are designed to process the feature information of the backbone architecture.Meanwhile,a pixel-level attention mask is designed to weight and fuse the feature,in order to improve the accuracy of interlacing degree detection.Then,based on the proposed enhancement algorithm for semantic information,a parallel detection method of interlacing degree is designed to achieve batch calculation for interlacing degree of multiple filament yarns.The algorithm is used to detect interlacing node,while connected domain analysis and mask extraction are used for parallel detection to extract independent regions of each filament yarn within the field.The parallel detection results are then fused to accurately obtain the interlacing degree detection results for each filament yarn.To validate the effectiveness of the proposed method,a synthetic filament yarn dataset is established using a self-developed interlacing degree detection device,and experimental verification is conducted.The results demonstrate that the proposed method can effectively improve the accuracy of detection.
作者
郑广智
彭添强
肖计春
吴高昌
李智
柴天佑
ZHENG Guang-Zhi;PENG Tian-Qiang;XIAO Ji-Chun;WU Gao-Chang;LI Zhi;CHAI Tian-You(State Key Laboratory of Synthetical Automation for Process Industries,Northeastern University,Shenyang 110819;School of Mechanical Engineering and Automation,Northeastern University,Shenyang 110819;National Engineering Research Center of Metallurgy Automation,Shenyang 110819)
出处
《自动化学报》
EI
CAS
CSCD
北大核心
2024年第10期1963-1976,共14页
Acta Automatica Sinica
基金
国家自然科学基金(62173077,61991404,62103092)资助。
关键词
网络度
图像语义分割
语义特征信息
并行检测策略
注意力机制
Interlacing degree
image semantic segmentation
semantic feature information
parallel detection strategy
attention mechanism
