leafDiseaseCNN: Insect detection and classification
based on improved convolutional neural network
To identify agricultural pests by a CNN system ...
Plant leaf diseases are closely related to people's daily life. Due to the wide variety of diseases, it is not only time-consuming and labor-intensive to identify and classify diseases by artificial eyes, but also easy to be misidentified with having a high error rate. Therefore, we proposed a deep learning-based method to identify and classify plant leaf diseases. The proposed method can take the advantages of neural network to extract the characteristics of disease parts, and thus to classify target disease areas. To address the issues of long training convergence time and too large model parameters, traditional convolutional neural network was improved by combining a structure of inception module, a squeeze-and-excitation (SE) module and an global pooling layer to identify diseases. Through the Inception structure, the feature data of the convolutional layer were fused in multi-scales to improve the accuracy on leaf disease dataset. Finally, the global average pooling layer was used instead of the fully connected layer to reduce the number of model parameters. Compared with some traditional convolutional neural networks, VGGNet, ResNet, Inception-v2, Inception-v3 and SENet models, our model yielded better performance and achieved an accuracy of 91.7% on the test data set. At the same time, the number of model parameters and training time have also been greatly reduced. The experimental classification on plant leaf diseases indicated that our method is feasible and effective.
Figure 1. Sample images of 10 leaf diseases. (1) Apple healthy (AH); (2) Apple Scab general (ASG); (3) Apple Scab serious (ASS); (4) Apple Frogeye Spot (AFS); (5) Cedar Apple Rust genera (CARG)l; (6) Cedar Apple Rust serious (CARS); (7) Cherry healthy (CH); (8) Cherry Powdery Mildew general (CPMG); (9) Cherry Powdery Mildew serious (CPMS); (10) Corn healthy (CH).
Figure 9. Visualization of feature map from each layer for a sample leaf. (1) conv1_1, (2) conv3_1, (3) conv5_1, (4) inception_ 1x1, (5) inception_ 3x3, (6) inception_ 5x5, (7) inception_ pool, (8) pool7.
Source codes: supplementary.
Database:from webdisc
Copyright @ 2004-2018 by Peng Chen
All Rights Reserved