基于多尺度特征融合的马铃薯疮痂病图像语义分割方法

doi:10.6040/j.issn.1672-3961.0.2024.232

山东大学学报 (工学版) ›› 2025, Vol. 55 ›› Issue (4): 1-8.doi: 10.6040/j.issn.1672-3961.0.2024.232

• 深度学习与视觉专题 •

基于多尺度特征融合的马铃薯疮痂病图像语义分割方法

吴秋兰¹,尚素雅^1,2,张家辉³,孙守鑫¹,张峰¹,周波^3,4*,高峥³,史文宠³

1.山东农业大学信息科学与工程学院, 山东泰安 271018;2.山东省葡萄研究院, 山东济南 250100;3.山东农业大学生命科学学院, 山东泰安 271018;4.真核生物科技(山东)有限公司, 山东泰安 271000

发布日期:2025-08-31
作者简介:吴秋兰(1975— ),女,山东济宁人,教授,硕士生导师,博士,主要研究方向为农业信息化. E-mail:zxylsg@sdau.edu.cn. *通信作者简介:周波(1972— ),男,山东济南人,教授,博士生导师,博士,主要研究方向为马铃薯疮痂病和根结线虫生物防治体系. E-mail:Zhoubo2798@163.com
基金资助:
国家自然科学基金资助项目(60903098);山东大学微生物改造技术国家重点实验室开放课题资助项目(M2023-07);山东省自然科学基金面上资助项目(ZR2022MC109);宁夏回族自治区重点研发计划资助项目(2023BCF01015)

Semantic segmentation method for potato common scab images based on multiscale feature fusion

WU Qiulan¹, SHANG Suya^1,2, ZHANG Jiahui³, SUN Shouxin¹, ZHANG Feng¹, ZHOU Bo^3,4*, GAO Zheng³, SHI Wenchong³

WU Qiulan¹, SHANG Suya^{1, 2}, ZHANG Jiahui³, SUN Shouxin¹, ZHANG Feng¹, ZHOU Bo^{3, 4*}, GAO Zheng³, SHI Wenchong³(1. School of Information Science and Engineering, Shandong Agricultural University, Taian 271018, Shandong, China;
2. Shandong Academy of Grape, Jinan 250100, Shandong, China;
3. School of Life Sciences, Shandong Agricultural University, Taian 271018, Shandong, China;
4. Eukaryote Biotechnology(Shandong)Co., Ltd., Taian 271000, Shandong, China

Published:2025-08-31

摘要/Abstract

摘要： 为了精确分割马铃薯疮痂病斑,提出一种名为MSFF-UNet的语义分割模型。在模型解码器向上融合的同时进行特征增强,通过进行卷积和归一化操作,增强区分不同尺寸病斑的生长状况,也可增加多维度的特征融合功能,对解码器的高层次数据强化特征提取后与低层次数据进行特征融合,以捕获不同尺度下的马铃薯或疮痂病斑的语义数据。结果表明,改进后的马铃薯疮痂病图像语义分割模型精确率、类别平均像素准确率、平均交并比分别为93.90%、93.51%、87.72%,能够较准确地分割马铃薯与疮痂病斑。

关键词: 语义分割, 特征融合, UNet, 马铃薯, 疮痂病

Abstract: To precisely segment potato common scab lesions, a semantic segmentation model named MSFF-UNet was proposed. During the decoder's upward fusion process, feature enhancement was implemented through convolution and normalization operations to improve the differentiation of growth conditions in lesions of varying sizes. Additionally, multi-dimensional feature fusion capability was incorporated, where enhanced feature extraction from the decoder's high-level data was performed and subsequently fused with low-level data to capture semantic information of potatoes or common scab lesions at different scales. The results demonstrated that the improved semantic segmentation model achieved 93.90% precision, 93.51% mean class pixel accuracy, and 87.72% mean intersection over union, effectively enabling accurate segmentation of potatoes and common scab lesions.

Key words: semantic segmentation, feature fusion, UNet, potato, common scab

中图分类号:

TP391

吴秋兰,尚素雅,张家辉,孙守鑫,张峰,周波,高峥,史文宠. 基于多尺度特征融合的马铃薯疮痂病图像语义分割方法[J]. 山东大学学报 (工学版), 2025, 55(4): 1-8.

WU Qiulan, SHANG Suya, ZHANG Jiahui, SUN Shouxin, ZHANG Feng, ZHOU Bo, GAO Zheng, SHI Wenchong. Semantic segmentation method for potato common scab images based on multiscale feature fusion[J]. Journal of Shandong University(Engineering Science), 2025, 55(4): 1-8.

参考文献

[1] WANG Z J, LIU H, ZENG F K, et al. Potato processing industry in China: current scenario, future trends and global impact[J]. Potato Research, 2023, 66(2): 543-562.
[2] 杨冰, 平原, 杜春梅. 马铃薯疮痂病的致病机制及防治研究进展[J]. 中国农学通报, 2021, 37(18): 131-137. YANG Bing, PING Yuan, DU Chunmei. Pathogenic mechanism and control method of potato scab: research progress[J]. Chinese Agricultural Science Bulletin, 2021, 37(18): 131-137.
[3] 袁东华, 王承芳, 干华磊, 等. 解淀粉芽孢杆菌制剂与铜制剂联合防治马铃薯疮痂病研究[J]. 生物灾害科学, 2022, 45(2): 165-168. YUAN Donghua, WANG Chengfang, GAN Hualei, et al. Combined control of potato scab with bacillus amyloliquefaciens preparation and copper preparation[J]. Biological Disaster Science, 2022, 45(2): 165-168.
[4] LI Y T, LIU J Y, DÍAZ-CRUZ G, et al. Virulence mechanisms of plant-pathogenic Streptomyces species: an updated review[J]. Microbiology, 2019, 165(10): 1025-1040.
[5] 夏善勇, 牛志敏, 李庆全, 等. 马铃薯疮痂病菌及防控手段研究进展[J]. 中国瓜菜, 2022, 35(8): 12-17. XIA Shanyong, NIU Zhimin, LI Qingquan, et al. Disease caused by streptomyces spp. and its control measures[J]. China Cucurbits and Vegetables, 2022, 35(8): 12-17.
[6] ZHANG K L, WANG S X, HU Y H, et al. Evaluation method of potato storage external defects based on improved U-net[J]. Agronomy, 2023, 13(10): 2503-2520.
[7] AL-ADHAILEH M H, VERMA A, ALDHYANI T H H, et al. Potato blight detection using fine-tuned CNN architecture[J]. Mathematics, 2023, 11(6): 1516.
[8] SAMATHA B, KUMAR G R, RAO P A, et al. IoT-enabled potato diseases prediction using deep learning[C] //2023 International Conference on Innovative Data Communication Technologies and Application(ICIDCA).Uttarakhand,India,IEEE, 2023: 878-883.
[9] 张菁. 基于遗传蛙跳神经网络的马铃薯病斑图像分割研究[D]. 兰州: 甘肃农业大学, 2018. ZHANG Jing. The image segmentation of potato diseased spots based on genetic frog leaping neural network[D]. Lanzhou:Gansu Agricultural University, 2018.
[10] OPPENHEIM D, SHANI G, ERLICH O, et al. Using deep learning for image-based potato tuber disease detection[J]. Phytopathology, 2019, 109(6):1083-1087.
[11] 祁雁楠. 基于机器视觉的马铃薯疮痂检测方法[D].北京: 中国农业机械化科学研究院, 2019. QI Yannan. Detection method of potato scab based on machine vision[D]. Beijing: Chinese Academy of Agricultural Mechanization Science, 2019.
[12] CHOWDARY G J. Impact of machine learning models in pneumonia diagnosis with features extracted from chest X-rays using VGG16[J]. Turkish Journal of Computer and Mathematics Education(TURCOMAT), 2021, 12(5): 1521-1530.
[13] LEE H, LIM S. PU-MFA: point cloud up-sampling via multi-scale features attention[J]. Sensors, 2022, 22(23): 9308-9319.
[14] RUSSELL B C, TORRALBA A, MURPHY K P, et al. LabelMe: a database and web-based tool for image annotation[J]. International Journal of Computer Vision, 2008, 77(1): 157-173.
[15] 甘肃省粮食作物栽培标准化技术委员会. 马铃薯疮痂病测报技术规程: DB62/T 4080—2019[S]. [S.l.] :[s.n.] , 2019.
[16] 马丹丹, 关欢欢, 李寿如, 等. 马铃薯疮痂病菌在植株和田间的分布与动态分析[J]. 植物病理学报, 2022, 52(1): 61-67. MA Dandan, GUAN Huanhuan, LI Shouru, et al. Analysis of Streptomyces scabies distribution in plants and population dynamics in the field[J] Acta Phytopathologica Sinica, 2022, 52(1): 61-67.
[17] 王鹏程, 金光辉, 张春雨, 等. 不同生防菌剂组合及施用方式对马铃薯疮痂病的防治效果及促生作用[J]. 西南农业学报, 2022, 35(4): 797-803. WANG Pengcheng, JIN Guanghui, ZHANG Chunyu, et al. Biological control and growth promoting effect of potato common scab with different biocontrol agents and application method[J]. Southwest China Journal of Agricultural Sciences, 2022, 35(4): 797-803.
[18] DUAN C T, DING J H, CHEN S Y, et al. Temporal effective batch normalization in spiking neural networks[J]. Advances in Neural Information Processing Systems, 2022, 35: 34377-34390.
[19] SAHITO F, PAN Z W, SAHITO F, et al. Transpose convolution based model for super-resolution image reconstruction[J]. Applied Intelligence, 2023, 53(9): 10574-10584.
[20] 陈满, 金诚谦, 莫恭武,等. 基于改进U-Net模型的小麦收获含杂率在线检测方法[J]. 农业机械学报, 2023, 54(2): 73-82. CHEN Man, JIN Chengqian, MO Gongwu, et al. Online detection method of impurity rate in wheat mechanized harvesting based on improved U-Net model[J].Transactions of the Chinese Society for Agricultural Machinery, 2023, 54(2): 73-82.
[21] BOUSIAS ALEXAKIS E, ARMENAKIS C. Evaluation of UNet and UNet++ architectures in high resolution image change detection applications[J]. The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 2020, XLIII-B3-2020: 1507-1514.
[22] SARITURK B, BAYRAM B, DURAN Z D, et al. Feature extraction from satellite images using segnet and fully convolutional networks(FCN)[J]. International Journal of Engineering and Geosciences, 2020, 5(3): 138-143.
[23] YU L J, ZENG Z X, LIU A, et al. A lightweight complex-valued DeepLabv3+ for semantic segmentation of PolSAR image[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2022, 15: 930-943.

多维度评价

Viewed

Full text

Abstract

Cited

Shared

Discussed

基于多尺度特征融合的马铃薯疮痂病图像语义分割方法

Semantic segmentation method for potato common scab images based on multiscale feature fusion

PDF (PC)

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

多维度评价

本文评价

推荐阅读 0

[1]	刘全金,嵇文,胡浪涛,黄汇磊,杨瑞,李翔,高泽文,魏本征. 基于双解码器的医学图像分割模型[J]. 山东大学学报 (工学版), 2024, 54(6): 8-18.
[2]	杨巨成, 魏峰, 林亮, 贾庆祥, 刘建征. 驾驶员疲劳驾驶检测研究综述[J]. 山东大学学报 (工学版), 2024, 54(2): 1-12.
[3]	李明键,李卫军,王海荣. 融合词汇信息与GlobalPointer的实体识别[J]. 山东大学学报 (工学版), 2024, 54(1): 91-99.
[4]	迟云浩,杨璐,郭杰,郝凡昌,聂秀山. 基于注意力特征融合网络的手指静脉图像质量评价方法[J]. 山东大学学报 (工学版), 2023, 53(6): 56-62.
[5]	王碧瑶,韩毅,崔航滨,刘毅超,任铭然,高维勇,陈姝廷,刘嘉巍,崔洋. 基于图像的道路语义分割检测方法[J]. 山东大学学报 (工学版), 2023, 53(5): 37-47.
[6]	刘行,杨璐,郝凡昌. 基于多特征融合的手指静脉图像检索方法[J]. 山东大学学报 (工学版), 2023, 53(2): 118-126.
[7]	宋佳芮,陈艳平,王凯,黄瑞章,秦永彬. 基于Affix-Attention的命名实体识别语义补充方法[J]. 山东大学学报 (工学版), 2023, 53(2): 70-76.
[8]	李旭涛,杨寒玉,卢业飞,张玮. 基于深度学习的遥感图像道路分割[J]. 山东大学学报 (工学版), 2022, 52(6): 139-145.
[9]	尹旭,刘兆英,张婷,李玉鑑. 基于弱监督和半监督学习的红外舰船分割方法[J]. 山东大学学报 (工学版), 2022, 52(2): 99-106.
[10]	曹春红,段鸿轩,曹玲,张乐乐,胡凯,肖芬. 基于多级特征级联的遥感图像实时语义分割[J]. 山东大学学报 (工学版), 2021, 51(2): 19-25.
[11]	吴正健,木特力甫·马木提,吾尔尼沙·买买提,阿力木江·艾沙,库尔班·吾布力. 基于LTP和HOG纹理特征融合的中亚文档图像文种识别[J]. 山东大学学报 (工学版), 2021, 51(2): 115-121.
[12]	周杨浩,刘一帆,李瑮. 一种自动读取指针式仪表读数的方法[J]. 山东大学学报 (工学版), 2019, 49(4): 1-7.
[13]	张璞,刘畅,王永. 基于特征融合和集成学习的建议语句分类模型[J]. 山东大学学报 (工学版), 2018, 48(5): 47-54.
[14]	牟春倩,唐雁. 融合整体和局部信息的三维模型检索方法[J]. 山东大学学报(工学版), 2016, 46(6): 48-53.
[15]	王斌,常发亮,刘春生. 基于多特征融合的交通标志分类[J]. 山东大学学报(工学版), 2016, 46(4): 34-40.