P&amp;IDSeg-Net: 一种轻量化图纸分割网络

王晨曦; 杨婷婷; 张涛; 邹雨辰; 邵会凯; 钟德星

doi:null

您当前的位置：

首页 >

文章列表页 >

P&IDSeg-Net: 一种轻量化图纸分割网络

更新时间：2025-11-18

- P&IDSeg-Net: 一种轻量化图纸分割网络
- P&IDSeg-Net: A Lightweight P&ID Segmentation Network
- 西安交通大学学报 2025年
- 作者机构：
  
  1.西安交通大学自动化科学与工程学院,陕西省西安市710049
  2.中国广核集团有限公司科技数字化部,广东省深圳市518000
  3.中广核智能科技（深圳）有限责任公司,广东省深圳市518000
- 作者简介：
- 基金信息：
  
  中央高校基本科研业务费(xzy012023061)
- DOI：
  中图分类号： TP39
- 收稿：2025-06-25，
  
  修回：2025-10-25，
  
  录用：2025-10-27，
- 稿件说明：
移动端阅览
王晨曦, 杨婷婷, 张涛, 等. P&IDSeg-Net: 一种轻量化图纸分割网络[J/OL]. 西安交通大学学报, 2025.

WANG Chenxi, YANG Tingting, ZHANG Tao, et al. P&IDSeg-Net: A Lightweight P&ID Segmentation Network[J/OL]. JOURNAL OF XI’AN JIAOTONG UNIVERSITY, 2025.
王晨曦, 杨婷婷, 张涛, 等. P&IDSeg-Net: 一种轻量化图纸分割网络[J/OL]. 西安交通大学学报, 2025. DOI：

WANG Chenxi, YANG Tingting, ZHANG Tao, et al. P&IDSeg-Net: A Lightweight P&ID Segmentation Network[J/OL]. JOURNAL OF XI’AN JIAOTONG UNIVERSITY, 2025. DOI：

摘要

为了提高工业管道与仪表图（P&ID）智能分割的性能，提出了一种轻量化图纸分割网络P&IDSeg-Net。首先，设计了通道减半的非局部注意力模块，在深层特征空间建立全局依赖，解决了图纸中管道与仪器的结构混淆问题；其次，用特征加法融合代替通道拼接进行跳跃连接，进一步减少参数量，适应小规模数据集；然后，基于P&ID中管道区域的低灰度特性，构建了视觉提示词启发的灰度掩膜加权损失，使网络聚焦前景关键区域。最后，收集和手工标注了P&ID_Pipe数据集，并在该数据集上进行了大量的实验。结果表明：所提出的P&IDSeg-Net网络仅需21.81M参数，轻量化优势突出，同时平均交并比（MioU）达到77.40%，相比于主流的分割方法，能够获得更优的性能。

Abstract

To enhance the intelligent segmentation performance of Piping and Instrumentation Diagrams (P&ID)

a lightweight segmentation network named P&IDSeg-Net is proposed in this paper. First

a half-channel Non-local Attention Module is designed to establish global dependencies in deep feature space

addressing the structural ambiguities between pipelines and instruments. Second

feature addition fusion is employed instead of channel concatenation for skip connections to further reduce parameter count

making it suitable for small-scale datasets. Then

leveraging the low-grayscale characteristics of pipeline regions in P&IDs

a visual prompt inspired grayscale mask weighted loss is introduced to focus the network on key foreground regions. The P&ID_Pipe dataset is collected and manually annotated

and extensive experiments are conducted on it. The results demonstrate that the proposed P&IDSeg-Net requires only 21.81 million parameters

highlighting its lightweight advantages. Simultaneously

it achieves a Mean Intersection over Union (MIoU) of 77.40%

outperforming mainstream segmentation methods.

关键词

Keywords

references

Furuta M , Kase N , Emori S . Segmentation and recognition of symbols for handwritten piping and instrument diagram [C ] . Seventh International Conference on Pattern Recognition , Montreal, Que., Canada , 1984 : 626 – 629 .

Ishii M , Ito Y , Yamamoto M , et al . An automatic recognition system for piping and instrument diagrams [J ] . Systems and Computers in Japan , 1989 , 20 ( 3 ): 32 – 46 .

Elyan E , Garcia CM , Jayne C . Symbols Classification in Engineering Drawings [C ] . 2018 International Joint Conference on Neural Networks (IJCNN), Rio de Janeiro, Brazil ,, 2018 : 1 – 8 .

Kuchuganov V , Kuchuganov A , Kasimov D . Clustering Algorithm for a Set of Machine Parts on the Basis of Engineering Drawings [J ] . Programming and Computer Software , 2020 , 46 ( 1 ): 25 – 34 .

Moreno-García C , Elyan E , Jayne C . New trends on digitisation of complex engineering drawings [J ] . Neural Computing & Applications , 2019 , 31 ( 6 ): 1695 – 1712 .

范家斌 , 王宝会 , 陈继轩 . 基于文本-图像多模态融合的变电所布局图纸图符检测方法 [J ] . 计算机科学 , 2025 : 1 – 16 .

章喻龙 , 罗壮强 , 石凌峰 . 基于YOLOX算法的工程图纸标题栏识别 [J ] . 数字技术与应用 , 2024 , 42 ( 03 ): 198 – 202 .

侯凯 , 梅诗妍 . 基于特征提取和模板匹配的电网工程图纸字符识别技术 [J ] . 沈阳工业大学学报 , 2025 , 47 ( 02 ): 176 – 182 .

Rica E , Alvarez S , Serratosa F . Group of components detection in engineering drawings based on graph matching [J ] . Engineering Applications of Artificial Intelligence , 2021 , 104 : 1 – 10 .

Xie L , Lu Y , Furuhata T , et al . Graph neural network-enabled manufacturing method classification from engineering drawings [J ] . Computers in Industry , 2022 , 142 : 1 – 10 .

汪伟 . 基于人工智能的机械图纸图像切割技术的研究 [J ] . 精密制造与自动化 , 2025 , ( 01 ): 23 – 27 .

Zhang W , Joseph J , Yin Y , et al . Component segmentation of engineering drawings using Graph Convolutional Networks [J ] . Computers in Industry , 2023 , 147 : 1 – 12 .

OTSU N . THRESHOLD SELECTION METHOD FROM GRAY-LEVEL HISTOGRAMS [J ] . IEEE Transactions on Systems Man and Cybernetics , 1979 , 9 ( 1 ): 62 – 66 .

Hojjatoleslami S , Kittler J . Region growing: A new approach [J ] . IEEE Transactions on Image Processing , 1998 , 7 ( 7 ): 1079 – 1084 .

Long J , Shelhamer E , Darrell T . Fully Convolutional Networks for Semantic Segmentation [C ] . IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , Boston, MA , 2015 : 3431 – 3440 .

Noh H , Hong S , Han B . Learning Deconvolution Network for Semantic Segmentation [C ] . IEEE International Conference on Computer Vision , Santiago, CHILE , 2015 : 1520 – 1528 .

Ronneberger O , Fischer P , Brox T . U-Net: Convolutional Networks for Biomedical Image Segmentation [C ] . 18th International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI) , Munich, GERMANY , 2015 : 234 – 241 .

Chen L , Papandreou G , Kokkinos I , et al . DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs [J ] . IEEE Transactions on Pattern Analysis and Machine Intelligence , 2018 , 40 ( 4 ): 834 – 848 .

Zhao H , Shi J , Qi X , et al . Pyramid Scene Parsing Network [C ] . 30th IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , Honolulu, HI , 2017 : 6230 – 6239 .

Liu Z , Lin Y , Cao Y , et al . Swin Transformer: Hierarchical Vision Transformer using Shifted Windows [C ] . 18th IEEE/CVF International Conference on Computer Vision (ICCV), Electr Network , 2021 : 9992 – 10002 .

Xie E , Wang W , Yu Z , et al . SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers [C ] . 35th Annual Conference on Neural Information Processing Systems (NeurIPS), Electr Network , 2021 : 12077 – 12090 .

Kirillov A , Mintun E , Ravi N , et al . Segment anything [C ] . 2023 IEEE/CVF International Conference on Computer Vision (ICCV) , Paris, France , 2023 : 3992 – 4003 .

Ravi N , Gabeur V , Hu Y-T , et al . SAM 2: Segment Anything in Images and Videos [C ] . The Thirteenth International Conference on Learning Representations , (ICLR), Singapore , 2025 : 1 – 42 .

Chen T , Zhu L , Ding C , et al . SAM-Adapter: Adapting Segment Anything in Underperformed Scenes [C ] . International Conference on Computer Vision , (ICCV), Paris, France , 2023 : 3359 – 3367 .

Wang X , Girshick R , Gupta A , et al . Non-local Neural Networks [C ] . 31st IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , Salt Lake City, UT , 2018 : 7794 – 7803 .

Jia M , Tang L , Chen B , et al . Visual Prompt Tuning [C ] . 17th European Conference on Computer Vision (ECCV), Tel Aviv, ISRAEL , 2022 : 709 – 727 .

Mei J , Li R , Gao W , et al . CoANet: Connectivity Attention Network for Road Extraction From Satellite Imagery [J ] . IEEE Transactions on Image Processing , 2021 , 30 : 8540 – 8552 .

Chen J , Lu Y , Yu Q , et al . TransUNet: Transformers Make Strong Encoders for Medical Image Segmentation [J ] . CoRR , 2021 , abs/2102. 04306 : 1 – 13 .

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

P&IDSeg-Net：一种面向管道仪器图的轻量化分割网络

采用机器学习的大规模并行程序I/O性能建模与预测方法

电动车辆间接热系统快速除湿响应迟滞机理及优化控制策略

融合物理信息与深度算子网络的变压器绕组热-流场快速计算方法

冷库水平空气幕热湿阻隔性能数值模拟研究

P&IDSeg-Net: 一种轻量化图纸分割网络

P&IDSeg-Net: A Lightweight P&ID Segmentation Network

DOI：

摘要

Abstract

关键词

Keywords

references