Recognition of Wheat Leaf Diseases Using Lightweight Convolutional Neural Networks against Complex Backgrounds
Wheat leaf diseases are considered to be the foremost threat to wheat yield. In the realm of crop disease detection, convolutional neural networks (CNNs) have emerged as important tools. The training strategy and the initial learning rate are key factors that impact the performance and training spee...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-10-01
|
| Series: | Life |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2075-1729/13/11/2125 |
| _version_ | 1797458646545727488 |
|---|---|
| author | Xiaojie Wen Minghao Zeng Jing Chen Muzaipaer Maimaiti Qi Liu |
| author_facet | Xiaojie Wen Minghao Zeng Jing Chen Muzaipaer Maimaiti Qi Liu |
| author_sort | Xiaojie Wen |
| collection | DOAJ |
| description | Wheat leaf diseases are considered to be the foremost threat to wheat yield. In the realm of crop disease detection, convolutional neural networks (CNNs) have emerged as important tools. The training strategy and the initial learning rate are key factors that impact the performance and training speed of the model in CNNs. This study employed six training strategies, including Adam, SGD, Adam + StepLR, SGD + StepLR, Warm-up + Cosine annealing + SGD, Warm-up + Cosine, and annealing + Adam, with three initial learning rates (0.05, 0.01, and 0.001). Using the wheat stripe rust, wheat powdery mildew, and healthy wheat datasets, five lightweight CNN models, namely MobileNetV3, ShuffleNetV2, GhostNet, MnasNet, and EfficientNetV2, were evaluated. The results showed that upon combining the SGD + StepLR with the initial learning rate of 0.001, the MnasNet obtained the highest recognition accuracy of 98.65%. The accuracy increased by 1.1% as compared to that obtained with the training strategy with a fixed learning rate, and the size of the parameters was only 19.09 M. The above results indicated that the MnasNet was appropriate for porting to the mobile terminal and efficient for automatically identifying wheat leaf diseases. |
| first_indexed | 2024-03-09T16:40:09Z |
| format | Article |
| id | doaj.art-77fb18111eb5468ab12a1ae32fde962a |
| institution | Directory Open Access Journal |
| issn | 2075-1729 |
| language | English |
| last_indexed | 2024-03-09T16:40:09Z |
| publishDate | 2023-10-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Life |
| spelling | doaj.art-77fb18111eb5468ab12a1ae32fde962a2023-11-24T14:52:16ZengMDPI AGLife2075-17292023-10-011311212510.3390/life13112125Recognition of Wheat Leaf Diseases Using Lightweight Convolutional Neural Networks against Complex BackgroundsXiaojie Wen0Minghao Zeng1Jing Chen2Muzaipaer Maimaiti3Qi Liu4Key Laboratory of the Pest Monitoring and Safety Control of Crops and Forests of the Xinjiang Uygur Autonomous Region, College of Agronomy, Xinjiang Agricultural University, Urumqi 830052, ChinaKey Laboratory of the Pest Monitoring and Safety Control of Crops and Forests of the Xinjiang Uygur Autonomous Region, College of Agronomy, Xinjiang Agricultural University, Urumqi 830052, ChinaKey Laboratory of the Pest Monitoring and Safety Control of Crops and Forests of the Xinjiang Uygur Autonomous Region, College of Agronomy, Xinjiang Agricultural University, Urumqi 830052, ChinaKey Laboratory of the Pest Monitoring and Safety Control of Crops and Forests of the Xinjiang Uygur Autonomous Region, College of Agronomy, Xinjiang Agricultural University, Urumqi 830052, ChinaKey Laboratory of the Pest Monitoring and Safety Control of Crops and Forests of the Xinjiang Uygur Autonomous Region, College of Agronomy, Xinjiang Agricultural University, Urumqi 830052, ChinaWheat leaf diseases are considered to be the foremost threat to wheat yield. In the realm of crop disease detection, convolutional neural networks (CNNs) have emerged as important tools. The training strategy and the initial learning rate are key factors that impact the performance and training speed of the model in CNNs. This study employed six training strategies, including Adam, SGD, Adam + StepLR, SGD + StepLR, Warm-up + Cosine annealing + SGD, Warm-up + Cosine, and annealing + Adam, with three initial learning rates (0.05, 0.01, and 0.001). Using the wheat stripe rust, wheat powdery mildew, and healthy wheat datasets, five lightweight CNN models, namely MobileNetV3, ShuffleNetV2, GhostNet, MnasNet, and EfficientNetV2, were evaluated. The results showed that upon combining the SGD + StepLR with the initial learning rate of 0.001, the MnasNet obtained the highest recognition accuracy of 98.65%. The accuracy increased by 1.1% as compared to that obtained with the training strategy with a fixed learning rate, and the size of the parameters was only 19.09 M. The above results indicated that the MnasNet was appropriate for porting to the mobile terminal and efficient for automatically identifying wheat leaf diseases.https://www.mdpi.com/2075-1729/13/11/2125convolutional neural networktraining strategyinitial learning ratewheat leaf diseases |
| spellingShingle | Xiaojie Wen Minghao Zeng Jing Chen Muzaipaer Maimaiti Qi Liu Recognition of Wheat Leaf Diseases Using Lightweight Convolutional Neural Networks against Complex Backgrounds Life convolutional neural network training strategy initial learning rate wheat leaf diseases |
| title | Recognition of Wheat Leaf Diseases Using Lightweight Convolutional Neural Networks against Complex Backgrounds |
| title_full | Recognition of Wheat Leaf Diseases Using Lightweight Convolutional Neural Networks against Complex Backgrounds |
| title_fullStr | Recognition of Wheat Leaf Diseases Using Lightweight Convolutional Neural Networks against Complex Backgrounds |
| title_full_unstemmed | Recognition of Wheat Leaf Diseases Using Lightweight Convolutional Neural Networks against Complex Backgrounds |
| title_short | Recognition of Wheat Leaf Diseases Using Lightweight Convolutional Neural Networks against Complex Backgrounds |
| title_sort | recognition of wheat leaf diseases using lightweight convolutional neural networks against complex backgrounds |
| topic | convolutional neural network training strategy initial learning rate wheat leaf diseases |
| url | https://www.mdpi.com/2075-1729/13/11/2125 |
| work_keys_str_mv | AT xiaojiewen recognitionofwheatleafdiseasesusinglightweightconvolutionalneuralnetworksagainstcomplexbackgrounds AT minghaozeng recognitionofwheatleafdiseasesusinglightweightconvolutionalneuralnetworksagainstcomplexbackgrounds AT jingchen recognitionofwheatleafdiseasesusinglightweightconvolutionalneuralnetworksagainstcomplexbackgrounds AT muzaipaermaimaiti recognitionofwheatleafdiseasesusinglightweightconvolutionalneuralnetworksagainstcomplexbackgrounds AT qiliu recognitionofwheatleafdiseasesusinglightweightconvolutionalneuralnetworksagainstcomplexbackgrounds |