| 682 | 23 | 75 |
| 下载次数 | 被引频次 | 阅读次数 |
为解决传统卷积神经网络模型训练时间长、参数量大、泛化能力弱等问题,提出了一种基于VGG-16的改进多尺度卷积神经网络模型。用一个叠加卷积层替换VGG-16模型的最后3×3×512卷积层,并进行批归一化处理,提高模型训练速度;用全局池化层替换全连接层,大大减少模型参数总量。利用Plant Village公共数据集(健康玉米叶片、灰斑病、锈病和叶枯病叶片)结合大田试验采集的玉米病害图像数据对改进后模型进行训练和测试,并与常见的传统卷积神经网络模型进行对比。结果表明,模型参数和收敛时间均小于传统卷积神经网络,单一背景下的平均分类识别准确率达99.31%,明显优于传统神经网络模型(VGG-16的90.89%、ResNet-50的93.60%、Inception-V3的94.23%、MobileNet-V2的93.83%和DenseNet-201的95.70%)。同时,利用大田复杂背景病害图片测试新模型的泛化性,识别准确率达98.44%,单张图片测试平均仅需0.25 s。
Abstract：In order to solve the problems of long training time, large amount of parameters, and weak generalization ability of traditional convolutional neural networks, this paper proposes an improved multi-scale convolutional network model based on VGG-16. The last 3×3×512 convolution layer of the VGG-16 model was replaced with a superimposed convolution layer for batch normalization to increase the model learning rate. In addition, the total amount of model parameters was reduced by replacing the fully connected layer with a global pooling layer. The public Plant Village data set(healthy maize leaves, gray spot disease, rust and leaf blight leaves)combined with field collected maize disease images were used to train and test the improved VGG model, and the results were compared with those of the traditional neural network models. The results show that the convergence time of the improved model is significantly shorter than that of the traditional convolutional neural networks, and the average classification accuracy reached 99.31%,which is much better than those of the traditional neural network models(VGG-16 90.89%, Resnet-50 93.60%,Inception-V3 94.23%,Mobilenet 93.83% and DenseNet-201 95.70%). Meanwhile, field disease images with complex backgrounds were used to test the proposed method, and the average recognition accuracy could reach 98.44%, taking only 0.25 seconds per image.
[1] 计甜甜，李泽彬，赵江东，等.基于颜色差异性的植物叶片病害图像分割方法[J].湖北农业科学，2018,57(18):94-97.
[2] 张会敏，谢泽奇，张善文，等.基于WT-Otsu算法的植物病害叶片图像分割方法[J].江苏农业科学，2017,45(18):194-196.
[3] 张开兴，吕高龙，贾浩，等.基于图像处理和BP神经网络的玉米叶部病害识别[J].中国农机化学报，2019,40(8):122-126.
[4] CAGLAYAN A,GUCLU O,CAN A B.A plant recognition approach using shape and color features in leaf images[C].International Conference on Image Analysis and Processing,2013:161-170.
[5] LEE K B,CHUNG K W,HONG K S.An implementation of leaf recognition system based on leaf contour and centroid for plant classification[J].Lecture Notes in Electrical Engineering,2013,214:109-116.
[6] MAINA C N.Vision-based model for maize leaf disease identification:A case study in Nyeri County [D].Strathmore University,2016:3-7.
[7] ZHANG S W,WANG H X,HUANG W Z,et al.Plant diseased leaf segmentation and recognition by fusion of superpixel,K-means and PHOG[J].Optik-International Journal for Light and Electron Optics,2018,157:866-872.
[8] 贾浩.基于计算机视觉的玉米叶部病害识别技术的研究[D].泰安：山东农业大学，2014.
[9] 顾博，邓蕾蕾，李巍，等.基于GrabCut算法的玉米病害图像识别方法研究[J].中国农机化学报，2019,40(11):143-149.
[10] KRIZHEVSKY A,SUTSKEVER I,HINTON G.ImageNet classification with deep convolutional neural networks[J].Advances in Neural Information Processing Systems,2017,60(6):84-90.
[11] SIMONYAN K,ZISSERMAN A.Very deep convolutional networks for large-scale image recognition[J].Computer Science,2014,71:149-156.
[12] HE K,ZHANG X Y,REN S Q,et al.Deep residual learning for image recognition[C]//Proceedings of the IEEE conference on computer vision and patternrecognition,2016:770-778.
[13] SZEGEDY C,VANHOUCKE V,IOFFE S,et al.Rethinking the inception architecture for computer vision[J].IEEE Conference on Computer Vision and Pattern Recognition,2016,39:2818-2826.
[14] HUANG G,LIU Z,LAURENS V,et al.Densely Connected Convolutional Networks[C]//IEEE Computer Society.IEEE Computer Society,2017:4700-4708.
[15] SANDLER M,HOWARD A,ZHUM L,et al.MobileNetV2:Inverted Residuals and Linear Bottlenecks[C]//IEEE/CVF Conference on Computer Vision and Pattern Recognition,2018:4510-4520.
[16] HOWARD A G,ZHU M,CHEN B,et al.MobileNets:Efficient convolutional neural networks for mobile vision applications[J] .Ar Xiv Preprint,2017:71:139-151.
[17] KAMAL K C,YIN Z,WU M,et al.Depthwise separable convolution architectures for plant disease classification[J].Computers and Electronics in Agriculture,2019,165:121-132.
[18] DARWISH A,EZZAT D,HASSANIEN A E .An optimized model based on convolutional neural networks and orthogonal learning particle swarm optimization algorithm for plant diseases diagnosis[J].Swarm and Evolutionary Computation,2020,52:47-58.
[19] GANE G,JEPR A P.Identification of plant leaf diseases using a nine-layer deep convolutional neural network[J].Computers and Electrical Engineering,2019,76:323-338.
[20] CHEN J D,CHEN J X,ZHANG D F,et al.Using deep transfer learning for image-based plant disease identification[J].Computers and Electronics in Agriculture,2020,173:168-171.
[21] 郑方梅.基于卷积神经网络的农作物病害图像识别研究[D].重庆：重庆师范大学，2019.
[22] 魏超，范自柱，张泓，等.基于深度学习的农作物病害检测[J].江苏大学学报(自然科学版),2019,40(2):190-196.
[23] LIN M,CHEN Q,YAN S.Network in network[J].Computer Science,2013,56:135-142.
[24] IOFFE S,SZEGEDY C.Batchnormalization:Accelerating deep network training by reducing internal covariate shift[C]//International conference on machine learning.PMLR,2015:448-456.
[25] HUGHES D P,SALATHEM.An open access repository of images on plant health to enable the development of mobile disease diagnostics[J].Computer Science,2015:1511:121-139.
[26] ROTHER C.GrabCut:Interactive foreground extraction using iterated graph cut[J].Acm Trans Graph,2004,23:309-314.
[27] 王丹.随机梯度下降算法研究[D].西安：西安建筑科技大学，2020.
[28] 崔义新.基于交叉熵的随机赋权网络[D].保定：河北大学，2017.
[29] SRIVASTAVA N,HINTON G E,ALEX K,et al.Dropout:a simple way to prevent neural networks from overfitting.[J].Journal of Machine Learning Research,2014,15(1):1929-1958.
[30] 葛恭豪.机器学习算法原理及效率分析[J].电子世界，2018(1):65-66.
[31] 张陶.基于学习率衰减的深度学习超参数优化方法的研究[D].武汉：华中师范大学，2020.
基本信息:
DOI：10.16445/j.cnki.1000-2340.20210824.001
中图分类号:S435.131;TP183;TP391.41
引用信息:
[1]王美娟,尹飞.卷积神经网络的多尺度改进及其在玉米病害症状识别中的应用[J].河南农业大学学报,2021,55(05):906-916.DOI:10.16445/j.cnki.1000-2340.20210824.001.
基金信息:
国家重点研发计划项目(2017YFD0301105)