基于改进 ShuffleNetV2 与模型压缩的椰子幼苗病害识别方法 研究
Journal: Agricultural Science DOI: 10.32629/as.v9i5.3970
Abstract
椰子幼苗易受叶片病害影响。随着深度学习发展,卷积神经网络已广泛应用于植物病 害识别。然而,现有方法仍存在多尺度特征提取不足及模型复杂度较高的问题。为此 ,本文以椰子幼苗病害为识别对象,提出一种基于改进ShuffleNetV2的椰子幼苗病害 识别方法,通过引入多尺度特征提取与融合以及注意力增强,有效提升模型对不同尺 度病斑与关键区域识别能力,最后基于L1范数结构化剪枝的模型压缩方法,实现精度 下降可控与模型规模的显著压缩。本文提出的方法与当前具有代表性的8个基线方法 对比,在各项评价指标上均取得了不同程度的提升,其中准确率、精确度、召回率和 F1值的最大提升分别达到8.75%、8.21%、8.01%和8.09%,FLOPs 和参数量分别最高降低约80.00%和90.43%,最后压缩后的模型在各对比方法中取得了 最优的综合表现,其准确率达到98.05%,仅较原始模型下降0.15%。同时模型参数量降 低至1.10M,FLOPs降低至0.70G,压缩率和加速率分别达到2.44×和2.43×,说明所提出 方法能够在模型压与性能保持之间取得更好的平衡。
Keywords
椰子幼苗病害；轻量化卷积神经网络；多尺度特征；注意力机制
Full Text
PDF - Viewed/Downloaded: 7 TimesReferences
[1] Gurr G M,Johnson A C,Ash G J,et al.Coconut lethal yello wing diseases: a phytoplasma threat to palms of global econo mic and social significance[J].Frontiers in plant science,2016, 7:1521.
[2] Sandler M, Howard A, Zhu M, et al. Mobilenetv2: Inverted residuals and linear bottlenecks[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2018: 4510-4520.
[3] Wang Q,Wu B,Zhu P,et al.ECA-Net:Efficient channel atte ntion for deep convolutional neural networks[C]//Proceedings of the IEEE/CVF conference on computer vision and pattern recognition.2020:11534- 11542.
[4] Ma N, Zhang X, Zheng H T, et al. Shufflenet v2: Practical guidelines for efficient cnn architecture design[C]//Proceed ings of the European conference on computer vision (ECCV). 2018:116-131.
[5] Howard A G, Zhu M, Chen B, et al. Mobilenets: Efficient convolutional neural networks for mobile vision applications [J].arXiv preprint arXiv:1704.04861,2017.
[6] Howard A,Sandler M,Chu G,et al.Searching for mobilene tv3[C]//Proceedings of the IEEE/CVF international conference on computer vision.2019:1314-1324.
[7] Tan M, Le Q. Efficientnet: Rethinking model scaling for convolutional neural networks[C]//International conference on machine learning.PMLR,2019:6105-6114.
[8] Hou Q,Zhou D, Feng J.Coordinate attention for efficie nt mobile network design[C]//Proceedings of the IEEE/CVF con ference on computer vision and pattern recognition. 2021: 13713-13722.
[9] Li Y,Hou Q,Zheng Z,et al. Large selective kernel network for remote sensing object detection[C]//Proceedings of the IEEE/CVF international conference on computer vision. 2023: 16794-16805.
[10] Hinton G, Vinyals O, Dean J. Distilling the knowledge in a neural network[J]. arXiv preprint arXiv:1503.02531, 2015.
[11] Li H, Kadav A, Durdanovic I, et al. Pruning filters for efficient convnets[J]. arXiv preprint arXiv:1608.08710,2016.
[12] He Y, Liu P, Wang Z, et al. Filter pruning via geometric median for deep convolutional neural networks acceleration [C]//Proceedings of the IEEE/CVF conference on computer visi on and pattern recognition.2019:4340-4349.
[13] Liu Y,Fan K, Wu D, et al. Filter pruning by quantifying feature similarity and entropy of feature maps[J].Neurocomp uting,2023,544:126297.
[14] Li J,Shao H,Zhai S,et al.A graphical approach for filt er pruning by exploring the similarity relation between feat ure maps[J].Pattern Recognition Letters,2023,166:69-75.
[15] Lin M, Ji R, Wang Y, et al. Hrank: Filter pruning using high-rank feature map[C]//Proceedings of the IEEE/CVF confer ence on computer vision and pattern recognition. 2020: 1529- 1538.
[16] Liu Z,Li J,Shen Z,et al.Learning efficient convolutio nal networks through network slimming[C]//Proceedings of the IEEE international conference on computer vision.2017:2736- 2744.
[17] Liu S, Huang D. Receptive field block net for accurate and fast object detection[C]//Proceedings of the European conference on computer vision(ECCV).2018:385-400.
[18] Yu F,Koltun V.Multi-scale context aggregation by dil ated convolutions[J]. arXiv preprint arXiv:1511.07122, 2015.
[19] Chen J,Kao S,He H,et al. Run, don't walk: chasing higher FLOPS for faster neural networks[C]//Proceedings of the IEEE/ CVF conference on computer vision and pattern recognition. 2023:12021- 12031.
[20] Lian S,Cai J, Zhao Y, et al. A cross-layer bidirectional measurement pruning for model compression guided by feat ure propagation[J].Information Sciences,2026:123216.
[21] Pal C,Karmakar S, Mukherjee I, et al. A lightweight and explainable CNN model for empowering plant disease diagnosis [J].Scientific Reports,2025,15(1):30720.
[22] Tonmoy M R, Hossain M M,Dey N,et al.MobilePlantViT:a Mobile-friendly hybrid ViT for generalized plant disease ima ge classification[J].arXiv preprint arXiv:2503.16628,2025.
[23] Thai H T,Le K H,Nguyen N L T.Towards sustainable agric ulture:A lightweight hybrid model and cloud-based collection of datasets for efficient leaf disease detection[J].Future Generation Computer Systems,2023,148:488-500.
[24] Liu Z,Lin Y,Cao Y, et al. Swin transformer: Hierarchical vision transformer using shifted windows[C]//Proceedings of the IEEE/CVF international conference on computer vision.2021: 10012-10022.
[25] Liu Y, Wu D, Zhou W, et al. EACP: An effective automatic channel pruning for neural networks[J]. Neurocomputing, 2023, 526:131-142.
[2] Sandler M, Howard A, Zhu M, et al. Mobilenetv2: Inverted residuals and linear bottlenecks[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2018: 4510-4520.
[3] Wang Q,Wu B,Zhu P,et al.ECA-Net:Efficient channel atte ntion for deep convolutional neural networks[C]//Proceedings of the IEEE/CVF conference on computer vision and pattern recognition.2020:11534- 11542.
[4] Ma N, Zhang X, Zheng H T, et al. Shufflenet v2: Practical guidelines for efficient cnn architecture design[C]//Proceed ings of the European conference on computer vision (ECCV). 2018:116-131.
[5] Howard A G, Zhu M, Chen B, et al. Mobilenets: Efficient convolutional neural networks for mobile vision applications [J].arXiv preprint arXiv:1704.04861,2017.
[6] Howard A,Sandler M,Chu G,et al.Searching for mobilene tv3[C]//Proceedings of the IEEE/CVF international conference on computer vision.2019:1314-1324.
[7] Tan M, Le Q. Efficientnet: Rethinking model scaling for convolutional neural networks[C]//International conference on machine learning.PMLR,2019:6105-6114.
[8] Hou Q,Zhou D, Feng J.Coordinate attention for efficie nt mobile network design[C]//Proceedings of the IEEE/CVF con ference on computer vision and pattern recognition. 2021: 13713-13722.
[9] Li Y,Hou Q,Zheng Z,et al. Large selective kernel network for remote sensing object detection[C]//Proceedings of the IEEE/CVF international conference on computer vision. 2023: 16794-16805.
[10] Hinton G, Vinyals O, Dean J. Distilling the knowledge in a neural network[J]. arXiv preprint arXiv:1503.02531, 2015.
[11] Li H, Kadav A, Durdanovic I, et al. Pruning filters for efficient convnets[J]. arXiv preprint arXiv:1608.08710,2016.
[12] He Y, Liu P, Wang Z, et al. Filter pruning via geometric median for deep convolutional neural networks acceleration [C]//Proceedings of the IEEE/CVF conference on computer visi on and pattern recognition.2019:4340-4349.
[13] Liu Y,Fan K, Wu D, et al. Filter pruning by quantifying feature similarity and entropy of feature maps[J].Neurocomp uting,2023,544:126297.
[14] Li J,Shao H,Zhai S,et al.A graphical approach for filt er pruning by exploring the similarity relation between feat ure maps[J].Pattern Recognition Letters,2023,166:69-75.
[15] Lin M, Ji R, Wang Y, et al. Hrank: Filter pruning using high-rank feature map[C]//Proceedings of the IEEE/CVF confer ence on computer vision and pattern recognition. 2020: 1529- 1538.
[16] Liu Z,Li J,Shen Z,et al.Learning efficient convolutio nal networks through network slimming[C]//Proceedings of the IEEE international conference on computer vision.2017:2736- 2744.
[17] Liu S, Huang D. Receptive field block net for accurate and fast object detection[C]//Proceedings of the European conference on computer vision(ECCV).2018:385-400.
[18] Yu F,Koltun V.Multi-scale context aggregation by dil ated convolutions[J]. arXiv preprint arXiv:1511.07122, 2015.
[19] Chen J,Kao S,He H,et al. Run, don't walk: chasing higher FLOPS for faster neural networks[C]//Proceedings of the IEEE/ CVF conference on computer vision and pattern recognition. 2023:12021- 12031.
[20] Lian S,Cai J, Zhao Y, et al. A cross-layer bidirectional measurement pruning for model compression guided by feat ure propagation[J].Information Sciences,2026:123216.
[21] Pal C,Karmakar S, Mukherjee I, et al. A lightweight and explainable CNN model for empowering plant disease diagnosis [J].Scientific Reports,2025,15(1):30720.
[22] Tonmoy M R, Hossain M M,Dey N,et al.MobilePlantViT:a Mobile-friendly hybrid ViT for generalized plant disease ima ge classification[J].arXiv preprint arXiv:2503.16628,2025.
[23] Thai H T,Le K H,Nguyen N L T.Towards sustainable agric ulture:A lightweight hybrid model and cloud-based collection of datasets for efficient leaf disease detection[J].Future Generation Computer Systems,2023,148:488-500.
[24] Liu Z,Lin Y,Cao Y, et al. Swin transformer: Hierarchical vision transformer using shifted windows[C]//Proceedings of the IEEE/CVF international conference on computer vision.2021: 10012-10022.
[25] Liu Y, Wu D, Zhou W, et al. EACP: An effective automatic channel pruning for neural networks[J]. Neurocomputing, 2023, 526:131-142.
Copyright © 2026 廖碧蓉, 李睿, 李雪宁, 王家豪, 黄闽英
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License