融合SKNet与MobilenetV3的芒果叶片病虫害分类方法

沈熠辉; 何惠彬; 陈小宇; 颜胜男

融合SKNet与MobilenetV3的芒果叶片病虫害分类方法

1.
福建医科大学文理艺术学院，福建　福州　350122
2.
福建省农业机械化研究所智能技术部，福建　福州　350005
3.
闽江学院计算机与控制工程学院，福建　福州　350103

基金项目: 福建省技术创新重点攻关及产业化项目（2023G015）

详细信息

作者简介:
沈熠辉（1991 —），男，硕士，助教，主要从事人工智能应用、农业数据处理研究，E-mail：1723378551@qq.com

中图分类号: TQ639.8;TP183
计量
- 文章访问数: 13
- HTML全文浏览量: 6
- PDF下载量: 0
- 被引次数: 0
出版历程
- 收稿日期: 2024-04-18
- 修回日期: 2024-05-05
- 网络出版日期: 2024-06-26

A Classification Method for Mango Leaf Diseases and Pests by Integrating SKNet and MobilenetV3

1.
College of Arts and Sciences, Fujian Medical University, Fuzhou　350122, China
2.
.Intelligent Technology Department of Fujian Agricultural Mechanization Research Institute, Fuzhou　350005, China
3.
School of Computer and Control Engineering, Minjiang University, Fuzhou　350103, China

摘要

摘要: 目的针对芒果叶片病虫害缺少数据集和识别准确率低的问题，提高芒果叶病虫害分类准确率。方法提出使用去噪扩散模型进行病虫害数据增强，同时联合SKNet与MobilenetV3模型的芒果叶片病虫害分类方法。首先使用去噪扩散模型对数据集进行扩充，再采用多尺度结构相似性指标对生成的病虫害图像与拍摄的病虫害图像之间的相似程度进行评估，接着对DDIM与DCGAN网络训练和生成效果进行比对。在MobilenetV3模型中，将SE注意力模块替换为SKNet模块进行构建网络模型。结果使用DDIM生成的所有类型的病虫害图像与拍摄的病虫害图像的MS-SSIM指标均大于0.63，且都高于DCGAN。相较于其他注意力模块，联合SKNet与MobilenetV3的分类效果最佳，在98%以上。对添加CA、CBAM、ECA注意力模块进行平滑类激活图可视化，对比其他注意力模块，使用SKNet注意力分布区域更为集中在病虫害叶片上。结论该方法在病虫害叶片检测上具有良好的应用前景，能提升病虫害识别效率与精度,减少检测成本，同时可应用于移动式或者嵌入式设备。
- 芒果叶片 /
- 扩散概率模型 /
- Mobilenet /
- Selective Kernel Networks
Abstract: : Objective To provide reference for the classification of mango leaf diseases, addressing the issues of lack of datasets and low recognition accuracy. Method This article proposes the use of denoising diffusion models for disease data augmentation, and innovatively proposes a mango leaf disease classification method that combines SKNet and MobilenetV3 models. Firstly, a denoising diffusion model was used to expand the dataset, and then a multi-scale structural similarity index was used to evaluate the similarity between the generated disease images and the captured disease images. Then, the training and generation effects of DDIM and DCGAN networks were compared. In the Mobilenet model, replace the SE attention module with the SKNet module to construct the network model. Results The MS-SSIM index of all types of disease images generated and captured using DDIM is greater than 0.63 and higher than DCGAN. Compared to other attention modules, the combination of SKNet and MobilenetV3 has the best classification performance, at over 98%. Visualize the smooth class activation diagram of adding CA, CBAM, and ECA attention modules. Compared with other attention modules, the attention distribution area of SKNet is more concentrated on diseased leaves. Conclusion This method has good application prospects in the detection of diseased leaves, which can improve the efficiency and accuracy of disease identification, reduce detection costs, and can be applied to mobile or embedded devices.
- mango leaf /
- diffusion probability model /
- Mobilenet /
- Selective Kernel Networks

HTML全文

图 1 病虫害叶片数据集样本

Figure 1. Sample of paint film defect dataset

下载: 全尺寸图片幻灯片

图 2 加噪与去噪示意图

Figure 2. Schematic diagram of adding and removing noise

下载: 全尺寸图片幻灯片

图 3 融合SKNet的bneck模型结构

注：⊕—元素相加运算；⊗—元素相乘运算

Figure 3. The bneck model structure integrating SKNet

Note: ⊕—Element-wise Addition; ⊗—Element-wise Multiplication

下载: 全尺寸图片幻灯片

图 4 学习率变化曲线

Figure 4. Change curve of Learning rate

下载: 全尺寸图片幻灯片

图 5 0-1000步学习率曲线

Figure 5. 0-1000 step Learning rate curve

下载: 全尺寸图片幻灯片

图 6 DDIM生成各类芒果叶片病虫害图

Figure 6. DDIM generates various mango leaf disease maps

下载: 全尺寸图片幻灯片

图 7 迭代次数0-80下数据增强前后损失值与准确率曲线图

Figure 7. Loss and Accuracy Curves Before and After Data Augmentation Over Epochs 0-80

下载: 全尺寸图片幻灯片

图 8 原始图像集（上）、DCGAN数据增强（中）、DDIM数据增强（下）分类混淆矩阵

Figure 8. Classification Confusion Matrix for Original Image Set(Top),DCGAN Data Augmentation(Middle), DDIM Data Augmentation (Bottom) Confusion matrix

下载: 全尺寸图片幻灯片

图 9 平滑类激活图（从上到下分别为添加SE、CA、CBAM、ECA、SKNet注意力模块的MobilenetV3模型）

Figure 9. Smooth Grad-CAM (from top to bottom are MobilenetV3 models with added SE, CA, CBAM, ECA, and SKNet attention modules, respectively).

下载: 全尺寸图片幻灯片

表 1 MobilenetV3模型结构

Table 1. MobilenetV3 model structure

输入尺寸 Input shape	操作算子 Operation	扩展尺寸 Expand size	输出通道 Output channel	SE模块 SE module	激活函数 Activation function	步长 Stride
224²	conv2d,3×3	—	16	—	HS	2
112²_×16	bneck,3×3	16	16	√	RE	2
56²_×16	bneck,3×3	72	24	—	RE	2
28²_×24	bneck,3×3	88	24	—	RE	1
28²_×24	bneck,5×5	96	40	√	HS	2
14²_×40	bneck,5×5	240	40	√	HS	1
14²_×40	bneck,5×5	240	40	√	HS	1
14²_×40	bneck,5×5	120	48	√	HS	1
14²_×48	bneck,5×5	144	48	√	HS	1
14²_×48	bneck,5×5	288	96	√	HS	2
7²_×96	bneck,5×5	576	96	√	HS	1
7²_×96	bneck,5×5	576	96	√	HS	1
7²_×96	conv2d,1×1	—	576	√	HS	1
7²_×576	pool,7×7	—	—	—	—	1
1²_×576	conv2d1×1,NBN	—	1280	—	HS	1
1²_×1024	conv2d1×1,NBN	—	7	—	—	1
注：conv2d—二维卷积；bneck—瓶颈模块；pool—池化层；NBN—不使用批量归一化；HS—硬切线激活函数；RE—修正线性单元；√—使用SE模块。 Note: conv2d—2D Convolution; bneck—Bottleneck Module; pool—Pooling Layer; NBN—No Batch Normalization; HS—Hard Swish Activation Function; RE—ReLU (Rectified Linear Unit); √—Uses SE Module

下载: 导出CSV

表 2 图片生成前后数据量比较

Table 2. Comparison of data volume before and after image generation

类型 Type	原始图数量 original dataset	增强后图像数量 Augmented dataset
炭疽病 Colletotrichum gloeosporioides	320	1800
细菌性溃疡 Xanthomonas campestris pv. mangiferaeindicae	324	1800
切叶象甲 Myllocerus viridanus	320	1800
枯萎病 Fusarium oxysporum	320	1800
癭蚊 Erosomyia mangiferae	544	1800
白粉病 Oidium mangiferae	320	1800
煤污病 Capnodium mangiferae	320	1800

下载: 导出CSV

表 3 DDIM与DCGAN网络模型训练指标对比

Table 3. DDIM vs. DCGAN Network Model Training Metrics Comparison

指标 Index	DDIM	DCGAN
模型大小 Model size/MB	117	89
训练时间 Training time /h	48	36
收敛速度(训练轮次) Convergence speed (epochs)	19	57
训练中损失函数的标准偏差 Standard deviation	0.05	0.15
总耗时（秒） Total time (seconds)	392.6	457.2

下载: 导出CSV

表 4 生成的病虫害图像与拍摄的病虫害图像MS-SSIM值

Table 4. MS-SSIM values of generated defect images and captured defect images

图像类型 Image type	MS-SSIM (DDIM)	MS-SSIM (DCGAN)
炭疽病 Colletotrichum gloeosporioides	0.6312	0.5992
细菌性溃疡 Xanthomonas campestris pv. mangiferaeindicae	0.7298	0.6912
切叶象甲 Myllocerus viridanus	0.6754	0.6413
枯萎病 Fusarium oxysporum	0.7123	0.6805
癭蚊 Erosomyia mangiferae	0.7459	0.6990
白粉病 Oidium mangiferae	0.7211	0.6853
煤污病 Capnodium mangiferae	0.6671	0.6396

下载: 导出CSV

表 5 引入注意力模块的MobilenetV3模型实验对比

Table 5. Experimental comparison of MobilenetV3 model with attention module introduction

算法 Algorithm	准确率 Accuracy/%	参数量 Params/M	乘加运算数 MACs/G
MobilenetV3(SE)	97.24	2.54	0.06
MobilenetV3+CA	95.75	2.18	0.06
MobilenetV3+CBAM	97.69	2.59	0.09
MobilenetV3+ECA	97.51	2.08	0.06
MobilenetV3+SKNet	98.21	2.58	0.06

下载: 导出CSV

参考文献(20)

[1]	杨芝霓. 芒果主要采后病害的病原鉴定及炭疽病的生物防治[D]. 南宁: 广西大学. YANG Z N. Pathogen identification of mango post havest disease and biological control of anthracnose[D]. Nanning: Guangxi University. (in Chinese)
[2]	秦丰, 刘东霞, 孙炳达, 等. 基于图像处理技术的四种苜蓿叶部病害的识别 [J]. 中国农业大学学报, 2016, 21(10):65−75. QIN F, LIU D X, SUN B D, et al. Recognition of four different alfalfa leaf diseases based on image processing technology [J]. Journal of China Agricultural University, 2016, 21(10): 65−75. (in Chinese)
[3]	柴阿丽, 李宝聚, 石延霞, 等. 基于计算机视觉技术的番茄叶部病害识别 [J]. 园艺学报, 2010, 37(9):1423−1430. CHAI A L, LI B J, SHI Y X, et al. Recognition of tomato foliage disease based on computer vision technology [J]. Acta Horticulturae Sinica, 2010, 37(9): 1423−1430. (in Chinese)
[4]	SHI B Z, ZHOU X L, QIN Z K, et al. Corn ear quality recognition based on DCGAN data enhancement and transfer learning[C]//The 4th International Conference on Electronics, Communications and Control Engineering. Seoul Republic of Korea. ACM, 2021.
[5]	AZIZI S, KORNBLITH S, SAHARIA C, et al. Synthetic data from diffusion models improves imagenet classification [J]. arXiv preprint arXiv:, 2304, 08466: 2023.
[6]	哈马友吉, 任万春, 张秤, 等. 基于轻量级网络MobileNet V2的二极管玻壳缺陷识别 [J]. 传感器与微系统, 2022, 41(4):153−155,160. HA M, REN W C, ZHANG C, et al. Defect recognition of diode glass shells based on lightweight network MobileNet V2 [J]. Transducer and Microsystem Technologies, 2022, 41(4): 153−155,160. (in Chinese)
[7]	王志强, 于雪莹, 杨晓婧, 等. 基于WGAN和MCA-MobileNet的番茄叶片病害识别 [J]. 农业机械学报, 2023, 54(5):244−252. WANG Z Q, YU X Y, YANG X J, et al. Tomato leaf diseases recognition based on WGAN and MCA-MobileNet [J]. Transactions of the Chinese Society for Agricultural Machinery, 2023, 54(5): 244−252. (in Chinese)
[8]	LI X, WANG W H, HU X L, et al. Selective kernel networks[J]. IEEE, 2020. DOI: 10.1109/CVPR.2019.00060.
[9]	AHMED S I, IBRAHIM M, NADIM M, et al. MangoLeafBD: A comprehensive image dataset to classify diseased and healthy mango leaves [J]. Data in Brief, 2023, 47: 108941. doi: 10.1016/j.dib.2023.108941
[10]	ZHANGQ, TAOM, CHENY. gDDIM: Generalizeddenoisingdiffusionimplicitmodels [J]. arXiv preprint arXiv:, 2206, 05564: 2022.
[11]	闫志浩, 周长兵, 李小翠. 生成扩散模型研究综述 [J]. 计算机科学, 2024, 51(1):273−283. YAN Z H, ZHOU C B, LI X C. Survey on generative diffusion model [J]. Computer Science, 2024, 51(1): 273−283. (in Chinese)
[12]	WANG Z, SIMONCELLI E P, BOVIK A C. Multiscale structural similarity for image quality assessment[C]//The Thrity-Seventh Asilomar Conference on Signals, Systems & Computers, 2003. Pacific Grove, CA, USA. IEEE, 2003: 1398-1402.
[13]	HOWARD A G, ZHU M L, CHEN B, et al. MobileNets: Efficient convolutional neural networks for mobile vision applications [J]. arXiv preprint arXiv:, 1704, 04861: 2017.
[14]	SANDLER M, HOWARD A, ZHU M L, et al. MobileNetV2: inverted residuals and linear bottlenecks[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. June 18-23, 2018. Salt Lake City, UT. IEEE, 2018: 4510-4520.
[15]	HOWARD A, SANDLER M, CHEN B, et al. Searching for MobileNetV3[C]//2019 IEEE/CVF International Conference on Computer Vision (ICCV). October 27-November 2, 2019. Seoul, Korea (South). IEEE, 2019: 1314-1324.
[16]	刘勇国, 高攀, 兰荻, 等. ECA-SKNet: 玉米单倍体种子的卷积神经网络识别模型 [J]. 电子科技大学学报, 2023, 52(6):866−871. LIU Y G, GAO P, LAN D, et al. ECA-SKNet: Convolutional neural network identification model for corn haploid seeds [J]. Journal of University of Electronic Science and Technology of China, 2023, 52(6): 866−871. (in Chinese)
[17]	何翔. 基于DCGANs的半片光伏组件电致发光图像增强技术 [J]. 应用光学, 2023, 44(2):314−322. doi: 10.5768/JAO202344.0202003 HE X. Electroluminescence image enhancement technology of half-cut photovoltaic module based on DCGANs [J]. Journal of Applied Optics, 2023, 44(2): 314−322. (in Chinese) doi: 10.5768/JAO202344.0202003
[18]	HOU Q B, ZHOU D Q, FENG J S. Coordinate attention for efficient mobile network design[C]//2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). June 20-25, 2021. Nashville, TN, USA. IEEE, 2021: 13713-13722.
[19]	WANG Q L, WU B G, ZHU P F, et al. ECA-net: Efficient channel attention for deep convolutional neural networks[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). June 13-19, 2020. Seattle, WA, USA. IEEE, 2020: 11534-11542.
[20]	WOO S, PARK J, LEE J Y, et al. CBAM: convolutional block attention module[M]//Computer Vision – ECCV 2018. Cham: Springer International Publishing, 2018: 3-19.