[1]张海婷,林 杨,金 曦,等.融合DBNet与LCNet-SVTR 的卷烟鉴别检验返样单自动识别方法[J].机械与电子,2026,44(07):61-70.
　ZHANG Haiting,LIN Yang,JIN Xi,et al.Automatic Recognition Method for Tobacco Identification and Inspection Return Samples Integrating DBNet and LCNet-SVTR[J].Machinery & Electronics,2026,44(07):61-70.
点击复制

融合DBNet与LCNet-SVTR 的卷烟鉴别检验返样单自动识别方法()
分享到：

《机械与电子》[ISSN:1001-2257/CN:52-1052/TH]

卷:
44
期数:
2026年07期
页码:
61-70
栏目:
智能检测
出版日期:
2026-07-25

文章信息/Info

Title:
Automatic Recognition Method for Tobacco Identification and Inspection Return Samples Integrating DBNet and LCNet-SVTR
文章编号:
1001-2257(2026)07-0061-10
作者:
张海婷1; 林 杨1; 金 曦1; 李秀梅1; 李书琼1; 袁庆霓2
1.贵州省烟草公司贵阳市公司,贵州 贵阳 550000;
2.贵州大学现代制造教育部重点实验室,贵州 贵阳 550025
Author(s):
ZHANG Haiting1; LIN Yang1; JIN Xi1; LI Xiumei1; LI Shuqiong1; YUAN Qingni2
(1.Guiyang City Company,Guizhou Tobacco Corporation,Guiyang 550000,China;
2.Key Laboratory of Advanced Manufacturing Technology of the Ministry of Education,Guizhou University,Guiyang 550025,China)
关键词:
深度学习; 文本检测; 轻量级卷积网络; 卷烟鉴别检验识别; 序列视觉
Keywords:
deep learning; text detection; lightweight convolutional network; tobacco identification and inspection recognition; sequence vision
分类号:
TP391
文献标志码:
A
摘要:
针对烟草行业传统人工处理鉴别检验返样单耗时费力、易出错,难以满足高效准确数据处理需求的问题,基于深度学习提出一种融合文本检测模型DBNet、轻量级卷积网络LCNet与序列视觉Transformer( SVTR)的卷烟鉴别检验返样单自动识别方法。该方法首先对返样单图像进行灰度归一化、对比度增强、尺度标准化及数据增强等预处理,以减弱扫描噪声、光照不均和印章干扰对识别性能的影响;随后,在文本检测阶段采用可微分二值化思想的文本检测模型DBNet,通过联合学习文本概率图与自适应阈值图,实现密集短文本区域的分割与定位;在文本识别阶段,构建融合LCNet与SVTR的LCNet-SVTR识别模型,以兼顾局部字符细节提取和全局上下文语义建模;最后,通过后处理解码输出返样单文本识别结果。实验结果表明,所提方法的识别准确率达到89%以上,归一化编辑距离达到98%以上,该方法为烟草行业的自动化文档处理提供了有效的解决方案。
Abstract:
Manual processing of tobacco identification and inspection return samples in the tobacco industry is time-consuming,labor-intensive,and error prone,failing to meet the demand for efficient and accurate data processing.To overcome these limitations,this paper proposes a deep learning-based automatic recognition method that integrates the differentiable binarization network (DBNet),a lightweight convolutional network LCNet,and the sequence vision Transformer (SVTR).The method first preprocesses return sample images through grayscale normalization,contrast enhancement,scale standardization,and data augmentation to suppress the adverse effects of scanning noise,uneven illumination,and seal interference on recognition performance. Subsequently,in the text detection stage,the DBNet model,which adopts the idea of differentiable binarization,is employed to jointly learn the text probability maps and an adaptive threshold map,thereby achieving the segmentation and localization of dense short text regions.In the text recognition stage,an LCNet-SVTR recognition model is constructed by combining LCNet and SVTR,which captures fine-grained local character detail while modeling global contextual semantics.Finally,the text recognition results of the return samples are output through post-processing decoding.Experimental results demonstrate that the proposed method achieves a recognition accuracy exceeding 89% and a normalized edit distance above 98%,offering an effective solution for automated document processing in the tobacco industry.

参考文献/References:

[1] Liao Minghui,Zou Zhaozhong,Wan Zhaoyi,et al.Real-time scene text detection with differentiable binarization and adaptive scale fusion[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2023,45(1): 919-931.

[2] Chen Xiaoxue,Jin Lianwen,ZhuYuanzhi,et al.Text recognition in the wild: a survey[J].ACM Computing Surveys,2021,54(2): 1-35.
[3] 王阳,李振东,杨观赐.基于深度学习的OCR文字识别在银行业的应用研究[J].计算机应用研究,2020,37(增刊2): 375-379.
[4] 王日花.基于深度学习的智能OCR识别关键技术及应用研究[J].邮电设计技术,2021(8):20-24.
[5] 姚鹏威.基于数字图像处理的表格识别[D].郑州: 郑州大学,2019.
[6] 连哲,殷雁君,云飞,等.基于深度学习的自然场景文本检测综述[J].计算机工程,2024,50(3): 16-27.
[7] 邓伟超,俞鑫春,朱建宝,等.基于DBNet-CST的文本检测模型[J].通信与信息技术,2025(1): 100-103.
[8] Xiang Liuqing,Wen Hanyu,Zhao Ming.Pill box text identification using DBNet-CRNN[J].International Journal of Environmental Research and Public Health,2023,20(5): 3881.
[9] 康智信,林姝,徐俊刚,等.面向智能巡检的电柜OCR识别方法研究[J].石油化工自动化,2025,61(4): 59-64.
[10] 郑金志,汲如意,张立波.基于Transformer解码的端到端场景文本检测与识别算法[J].通信学报,2023,44(5): 64-78.
[11] 冯宪伟,姚炜.基于双向长短期记忆网络与稀疏自注意力的票据文本识别方法[J].传感技术学报,2024,37(11): 1946-1951.
[12] Zhao Ming,Li Yalong,Zhang Chaolin,et al.SVTR-SRNet: a deep learning model for scene text recognition via SVTR framework and spatial reduction mechanism[J].Electronics,2024,13(23): 4756.
[13] Liao Minghui,Wan Zhaoyi,Yao Cong,et al.Real-time scene text detection with differentiable binarization[ C]∥Proceedings of the AAAI Conference on Artificial Intelligence,2020,34(7): 11474-11481.

相似文献/References:

[1]王骁贤,张保华,陆思良.基于连续小波变换和卷积神经网络的无刷直流电机故障诊断[J].机械与电子,2018,(06):29.
　WANG Xiaoxian,ZHANG Baohua,LU Siliang.Fault Diagnosis of Brushless Direct Current Motor Based on Continuous Wavelet Transform and Convolutional Neural Network[J].Machinery & Electronics,2018,(07):29.
[2]刘志宇,黄亦翔.基于深度学习和迁移学习的液压泵健康评估方法[J].机械与电子,2018,(09):67.
　LIU Zhiyu,HUANG Yixiang.Health Assessment for Hydraulic Pump Based on Deep Learning and Transfer Learning[J].Machinery & Electronics,2018,(07):67.
[3]肖倩宏,康 鹏,杜 江,等.深度学习在电网智能调控系统中应用研究[J].机械与电子,2021,(01):38.
　XIAO Qianhong,KANG Peng,DU Jiang,et al.Research on the Application of Deep Learning Theory in Power Grid Intelligent Dispatching[J].Machinery & Electronics,2021,(07):38.
[4]许 哲,张少帅,郭 璐,等.无人机深度学习去雾算法[J].机械与电子,2021,(04):13.
　XU Zhe,ZHANG Shaoshuai,GUO Lu,et al. Deep Learning Defogging Algorithm for UAV[J].Machinery & Electronics,2021,(07):13.
[5]齐爱玲１,李 琳１,朱亦轩２,等.基于融合特征的双通道CNN滚动轴承故障识别[J].机械与电子,2021,(05):15.
　QI Ailing,LI Lin,ZHU Yixuan,et al.Dual Channel CNN Bearing Fault Identification Based on Fusion Feature[J].Machinery & Electronics,2021,(07):15.
[6]徐先峰,郑少杰,赵 依,等.基于数据分解与重构的光伏发电功率超短期预测[J].机械与电子,2022,(04):20.
　XU Xianfeng,ZHENG Shaojie,ZHAO Yi,et al.Ultra-short-term Prediction of Photovoltaic Power Generation Based on Data Decomposition and Deconstruction[J].Machinery & Electronics,2022,(07):20.
[7]江 励,熊达明,汤健华,等.自然光线环境中的空间物体快速识别和定位算法研究[J].机械与电子,2022,(06):8.
　JIANG Li,XIONG Daming,TANG Jianhua,et al.Recognition and Positioning Algorithm of Space Objects in Natural Light Environment[J].Machinery & Electronics,2022,(07):8.
[8]王西志,管声启,张理博,等.基于视觉引导的工业棒材上料系统研究[J].机械与电子,2023,41(05):19.
　WANG Xizhi,GUAN Shengqi,ZHANG Libo,et al.Research on Industrial Bar Feeding System Based on Visual Guidance[J].Machinery & Electronics,2023,41(07):19.
[9]王 青,吕绪山,党 帅,等.基于深度学习的纱管识别方法研究[J].机械与电子,2023,41(12):20.
　WANG Qing,LYU Xushan,DANG Shuai,et al.Research on Yarn Bobbin Detection Method Based on Deep Learning[J].Machinery & Electronics,2023,41(07):20.
[10]姜越夫,王 青,吕绪山.改进 YOLOv5s 的纱管目标检测方法[J].机械与电子,2024,42(02):29.
　JIANG Yuefu,WANG Qing,LYU Xushan.Improved YOLOv5s Method for Yarn Tube Object Detection[J].Machinery & Electronics,2024,42(07):29.

备注/Memo

备注/Memo:
收稿日期:2026-04-20
基金项目:贵州省烟草公司贵阳市公司科技项目(黔烟筑科[2024] 4号)
作者简介:张海婷 (1983-),女,江苏徐州人,助理政工师,研究方向为烟草专卖案件数字化;袁庆霓 (1976-),女,贵州贵阳人,工学博士,教授,研究方向为智能机器人、数字化设计及制造、智能制造,通信作者,E-mail:qnyuan@gzu.edu.cn。
更新日期/Last Update: 2026-08-27