Pepper leaf disease recognition based on enhanced lightweight convolutional neural networks
Pepper leaf disease identification based on convolutional neural networks (CNNs) is one of the interesting research areas. However, most existing CNN-based pepper leaf disease detection models are suboptimal in terms of accuracy and computing performance. In particular, it is challenging to apply CN...
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2023-08-01
|
| Series: | Frontiers in Plant Science |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fpls.2023.1230886/full |
| _version_ | 1797748951880826880 |
|---|---|
| author | Min Dai Wenjing Sun Lixing Wang Md Mehedi Hassan Dorjoy Shanwen Zhang Hong Miao Hong Miao Liangxiu Han Xin Zhang Mingyou Wang |
| author_facet | Min Dai Wenjing Sun Lixing Wang Md Mehedi Hassan Dorjoy Shanwen Zhang Hong Miao Hong Miao Liangxiu Han Xin Zhang Mingyou Wang |
| author_sort | Min Dai |
| collection | DOAJ |
| description | Pepper leaf disease identification based on convolutional neural networks (CNNs) is one of the interesting research areas. However, most existing CNN-based pepper leaf disease detection models are suboptimal in terms of accuracy and computing performance. In particular, it is challenging to apply CNNs on embedded portable devices due to a large amount of computation and memory consumption for leaf disease recognition in large fields. Therefore, this paper introduces an enhanced lightweight model based on GoogLeNet architecture. The initial step involves compressing the Inception structure to reduce model parameters, leading to a remarkable enhancement in recognition speed. Furthermore, the network incorporates the spatial pyramid pooling structure to seamlessly integrate local and global features. Subsequently, the proposed improved model has been trained on the real dataset of 9183 images, containing 6 types of pepper diseases. The cross-validation results show that the model accuracy is 97.87%, which is 6% higher than that of GoogLeNet based on Inception-V1 and Inception-V3. The memory requirement of the model is only 10.3 MB, which is reduced by 52.31%-86.69%, comparing to GoogLeNet. We have also compared the model with the existing CNN-based models including AlexNet, ResNet-50 and MobileNet-V2. The result shows that the average inference time of the proposed model decreases by 61.49%, 41.78% and 23.81%, respectively. The results show that the proposed enhanced model can significantly improve performance in terms of accuracy and computing efficiency, which has potential to improve productivity in the pepper farming industry. |
| first_indexed | 2024-03-12T16:12:20Z |
| format | Article |
| id | doaj.art-456e09847502419781bb147fa902eadf |
| institution | Directory Open Access Journal |
| issn | 1664-462X |
| language | English |
| last_indexed | 2024-03-12T16:12:20Z |
| publishDate | 2023-08-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Plant Science |
| spelling | doaj.art-456e09847502419781bb147fa902eadf2023-08-09T11:47:03ZengFrontiers Media S.A.Frontiers in Plant Science1664-462X2023-08-011410.3389/fpls.2023.12308861230886Pepper leaf disease recognition based on enhanced lightweight convolutional neural networksMin Dai0Wenjing Sun1Lixing Wang2Md Mehedi Hassan Dorjoy3Shanwen Zhang4Hong Miao5Hong Miao6Liangxiu Han7Xin Zhang8Mingyou Wang9College of Mechanical Engineering, Yangzhou University, Yangzhou, ChinaCollege of Mechanical Engineering, Yangzhou University, Yangzhou, ChinaCollege of Mechanical Engineering, Yangzhou University, Yangzhou, ChinaCollege of Mechanical Engineering, Yangzhou University, Yangzhou, ChinaCollege of Mechanical Engineering, Yangzhou University, Yangzhou, ChinaCollege of Mechanical Engineering, Yangzhou University, Yangzhou, ChinaNanjing Institute of Agricultural Mechanization, Ministry of Agriculture and Rural Affairs, Nanjing, ChinaFaculty of Science and Engineering, Manchester Metropolitan University Manchester, Manchester, United KingdomFaculty of Science and Engineering, Manchester Metropolitan University Manchester, Manchester, United KingdomNanjing Institute of Agricultural Mechanization, Ministry of Agriculture and Rural Affairs, Nanjing, ChinaPepper leaf disease identification based on convolutional neural networks (CNNs) is one of the interesting research areas. However, most existing CNN-based pepper leaf disease detection models are suboptimal in terms of accuracy and computing performance. In particular, it is challenging to apply CNNs on embedded portable devices due to a large amount of computation and memory consumption for leaf disease recognition in large fields. Therefore, this paper introduces an enhanced lightweight model based on GoogLeNet architecture. The initial step involves compressing the Inception structure to reduce model parameters, leading to a remarkable enhancement in recognition speed. Furthermore, the network incorporates the spatial pyramid pooling structure to seamlessly integrate local and global features. Subsequently, the proposed improved model has been trained on the real dataset of 9183 images, containing 6 types of pepper diseases. The cross-validation results show that the model accuracy is 97.87%, which is 6% higher than that of GoogLeNet based on Inception-V1 and Inception-V3. The memory requirement of the model is only 10.3 MB, which is reduced by 52.31%-86.69%, comparing to GoogLeNet. We have also compared the model with the existing CNN-based models including AlexNet, ResNet-50 and MobileNet-V2. The result shows that the average inference time of the proposed model decreases by 61.49%, 41.78% and 23.81%, respectively. The results show that the proposed enhanced model can significantly improve performance in terms of accuracy and computing efficiency, which has potential to improve productivity in the pepper farming industry.https://www.frontiersin.org/articles/10.3389/fpls.2023.1230886/fulldeep convolutional neural networkscrop disease recognitionGoogLeNetreal-time recognitionlightweight neural networks |
| spellingShingle | Min Dai Wenjing Sun Lixing Wang Md Mehedi Hassan Dorjoy Shanwen Zhang Hong Miao Hong Miao Liangxiu Han Xin Zhang Mingyou Wang Pepper leaf disease recognition based on enhanced lightweight convolutional neural networks Frontiers in Plant Science deep convolutional neural networks crop disease recognition GoogLeNet real-time recognition lightweight neural networks |
| title | Pepper leaf disease recognition based on enhanced lightweight convolutional neural networks |
| title_full | Pepper leaf disease recognition based on enhanced lightweight convolutional neural networks |
| title_fullStr | Pepper leaf disease recognition based on enhanced lightweight convolutional neural networks |
| title_full_unstemmed | Pepper leaf disease recognition based on enhanced lightweight convolutional neural networks |
| title_short | Pepper leaf disease recognition based on enhanced lightweight convolutional neural networks |
| title_sort | pepper leaf disease recognition based on enhanced lightweight convolutional neural networks |
| topic | deep convolutional neural networks crop disease recognition GoogLeNet real-time recognition lightweight neural networks |
| url | https://www.frontiersin.org/articles/10.3389/fpls.2023.1230886/full |
| work_keys_str_mv | AT mindai pepperleafdiseaserecognitionbasedonenhancedlightweightconvolutionalneuralnetworks AT wenjingsun pepperleafdiseaserecognitionbasedonenhancedlightweightconvolutionalneuralnetworks AT lixingwang pepperleafdiseaserecognitionbasedonenhancedlightweightconvolutionalneuralnetworks AT mdmehedihassandorjoy pepperleafdiseaserecognitionbasedonenhancedlightweightconvolutionalneuralnetworks AT shanwenzhang pepperleafdiseaserecognitionbasedonenhancedlightweightconvolutionalneuralnetworks AT hongmiao pepperleafdiseaserecognitionbasedonenhancedlightweightconvolutionalneuralnetworks AT hongmiao pepperleafdiseaserecognitionbasedonenhancedlightweightconvolutionalneuralnetworks AT liangxiuhan pepperleafdiseaserecognitionbasedonenhancedlightweightconvolutionalneuralnetworks AT xinzhang pepperleafdiseaserecognitionbasedonenhancedlightweightconvolutionalneuralnetworks AT mingyouwang pepperleafdiseaserecognitionbasedonenhancedlightweightconvolutionalneuralnetworks |