# In-Field Citrus Disease Classification via Convolutional Neural Network from Smartphone Images

^{1}

^{2}

^{*}

## Abstract

**:**

## 1. Introduction

## 2. Dataset Construction

## 3. Methods

#### 3.1. Network-Based Transfer Learning

#### 3.2. Proposed 13-Layer Convolutional Neural Network (CNN13)

## 4. Experimental Results and Analysis

#### 4.1. Evaluation Metrics

#### 4.2. Experiments with OplusVNet with Different Frozen Mechanisms

#### 4.3. Comparison with State-of-the-Art Networks

#### 4.4. OplusVNet Network Performance Analysis

#### 4.4.1. Experimental Results of Proposed OplusVNet on Small Datasets

#### 4.4.2. Experimental Results of Different Network Models on Small Datasets

## 5. Conclusions

## Author Contributions

## Funding

## Institutional Review Board Statement

## Informed Consent Statement

## Data Availability Statement

## Conflicts of Interest

## References

- Khan, M.A.; Akram, T.; Sharif, M.; Awais, M.; Javed, K.; Ali, H.; Saba, T. CCDF: Automatic system for segmentation and recognition of fruit crops diseases based on correlation coefficient and deep CNN features. Comput. Electron. Agric.
**2018**, 155, 220–236. [Google Scholar] [CrossRef] - Moriya, E.A.S.; Imai, N.N.; Tommaselli, A.M.G.; Berveglieri, A.; Santos, G.H.; Soares, M.A.; Marino, M.; Reis, T.T. Detection and mapping of trees infected with citrus gummosis using UAV hyperspectral data. Comput. Electron. Agric.
**2021**, 188, 106298. [Google Scholar] [CrossRef] - Ngugi, L.C.; Abelwahab, M.; Abo-Zahhad, M. Recent advances in image processing techniques for automated leaf pest and disease recognition: A review. Inf. Process. Agric.
**2021**, 8, 27–51. [Google Scholar] [CrossRef] - Abdu, A.M.; Mokji, M.M.; Sheikh, U.U. Automatic vegetable disease identification approach using individual lesion features. Comput. Electron. Agric.
**2020**, 176, 105660. [Google Scholar] [CrossRef] - Nanehkaran, Y.A.; Zhang, D.; Chen, J.; Tian, Y.; Al-Nabhan, N. Recognition of plant leaf diseases based on computer vision. J. Ambient. Intell. Humaniz. Comput.
**2020**, 1–18. [Google Scholar] [CrossRef] - Jagan, K.; Balasubramanian, M.; Palanivel, S. Detection and Recognition of Diseases from Paddy Plant Leaf Images. Int. J. Comput. Appl.
**2016**, 144, 34–41. [Google Scholar] [CrossRef] - Hsiao, J.K.; Kang, L.W.; Chang, C.L.; Hsu, C.Y.; Chen, C.Y. Learning sparse representation for leaf image recognition. In Proceedings of the 2014 IEEE International Conference on Consumer Electronics, Taipei, Taiwan, 26–28 May 2014; pp. 209–210. [Google Scholar] [CrossRef]
- Zhang, S.; Wu, X.; You, Z.; Zhang, L. Leaf image based cucumber disease recognition using sparse representation classification. Comput. Electron. Agric.
**2017**, 134, 135–141. [Google Scholar] [CrossRef] - Liu, J.; Wang, X. Plant diseases and pests detection based on deep learning: A review. Plant Methods
**2021**, 17, 22. [Google Scholar] [CrossRef] - Chen, J.; Zhang, D.; Nanehkaran, Y. Identifying plant diseases using deep transfer learning and enhanced lightweight network. Multimed. Tools Appl.
**2020**, 79, 31497–31515. [Google Scholar] [CrossRef] - Chen, J.; Yin, H.; Zhang, D. A self-adaptive classification method for plant disease detection using GMDH-Logistic model. Sustain. Comput. Inform. Syst.
**2020**, 28, 100415. [Google Scholar] [CrossRef] - Chen, J.; Chen, J.; Zhang, D.; Sun, Y.; Nanehkaran, Y. Using deep transfer learning for image-based plant disease identification. Comput. Electron. Agric.
**2020**, 173, 105393. [Google Scholar] [CrossRef] - Chen, J.; Zhang, D.; Nanehkaran, Y.; Li, D. Detection of rice plant diseases based on deep transfer learning. J. Sci. Food Agric.
**2020**, 100. [Google Scholar] [CrossRef] - Haque, M.; Marwaha, S.; Deb, C.; Nigam, S.; Arora, A.; Hooda, K.; Soujanya, P.L.; Aggarwal, S.; Lall, B.; Kumar, M.; et al. Deep learning-based approach for identification of diseases of maize crop. Sci. Rep.
**2022**, 12, 6334. [Google Scholar] [CrossRef] [PubMed] - Kundu, N.; Rani, G.; Dhaka, V.S.; Gupta, K.; Nayak, S.C.; Verma, S.; Ijaz, M.F.; Woniak, M. IoT and Interpretable Machine Learning Based Framework for Disease Prediction in Pearl Millet. Sensors
**2021**, 21, 5386. [Google Scholar] [CrossRef] [PubMed] - Rehman, M.Z.U.; Ahmed, F.; Khan, M.A.; Tariq, U.; Jamal, S.S.; Ahmad, J.; Hussain, I. Classification of Citrus Plant Diseases Using Deep Transfer Learning. Comput. Mater. Contin.
**2022**, 70, 1401–1417. [Google Scholar] [CrossRef] - Nazari, K.; Ebadi, M.; Berahmand, K. Diagnosis of Alternaria disease and Leafminer pest on Tomato Leaves using Image Processing Techniques. J. Sci. Food Agric.
**2022**. [Google Scholar] [CrossRef] [PubMed] - Dhiman, P.; Kukreja, V.; Manoharan, P.; Kaur, A.; Kamruzzaman, M.M.; Dhaou, I.B.; Iwendi, C. A Novel Deep Learning Model for Detection of Severity Level of the Disease in Citrus Fruits. Electronics
**2022**, 11, 495. [Google Scholar] [CrossRef] - Mohanty, S.P.; Hughes, D.P.; Salath, M. Using Deep Learning for Image-Based Plant Disease Detection. Front. Plant Sci.
**2016**, 7, 1419. [Google Scholar] [CrossRef] - Sladojevic, S.; Arsenovic, M.; Anderla, A.; Culibrk, D.; Stefanovic, D. Deep Neural Networks Based Recognition of Plant Diseases by Leaf Image Classification. Comput. Intell. Neurosci.
**2016**, 2016, 3289801. [Google Scholar] [CrossRef][Green Version] - Tan, C.; Sun, F.; Kong, T.; Zhang, W.; Yang, C.; Liu, C. A Survey on Deep Transfer Learning. In Proceedings of the 27th International Conference on Artificial Neural Networks, Rhodes, Greece, 4–7 October 2018; pp. 270–279. [Google Scholar]
- Thenmozhi, K.; Reddy, U.S. Crop pest classification based on deep convolutional neural network and transfer learning. Comput. Electron. Agric.
**2019**, 164, 104906. [Google Scholar] [CrossRef] - Selvaraj, M.G.; Vergara, A.; Ruiz, H.; Safari, N.; Elayabalan, S.; Ocimati, W.; Blomme, G. AI-powered banana diseases and pest detection. Plant Methods
**2019**, 15, 92. [Google Scholar] [CrossRef] - Coulibaly, S.; Kamsu-Foguem, B.; Kamissoko, D.; Traore, D. Deep neural networks with transfer learning in millet crop images. Comput. Ind.
**2019**, 108, 115–120. [Google Scholar] [CrossRef] - Barman, U.; Choudhury, R.D.; Sahu, D.; Barman, G.G. Comparison of convolution neural networks for smartphone image based real time classification of citrus leaf disease. Comput. Electron. Agric.
**2020**, 177, 105661. [Google Scholar] [CrossRef] - Morteza, K.; Ezzatollah, A.A.A.; Ehsan, K. Citrus pests classification using an ensemble of deep learning models. Comput. Electron. Agric.
**2021**, 186, 106192. [Google Scholar] - Simonyan, K.; Zisserman, A. Very Deep Convolutional Networks for Large-Scale Image Recognition. arXiv
**2014**, arXiv:1409.1556. [Google Scholar] - Krizhevsky, A.; Sutskever, I.; Hinton, G.E. ImageNet Classification with Deep Convolutional Neural Networks. In Proceedings of the Annual Conference on Neural Information Processing Systems, Lake Tahoe, NV, USA, 3–8 December 2012; Volume 25. [Google Scholar]
- Krizhevsky, A.; Sutskever, I.; Hinton, G.E. ImageNet Classification with Deep Convolutional Neural Networks. Commun. ACM
**2017**, 60, 84–90. [Google Scholar] [CrossRef] - Srivastava, N.; Hinton, G.; Krizhevsky, A.; Sutskever, I.; Salakhutdinov, R. Dropout: A Simple Way to Prevent Neural Networks from Overfitting. J. Mach. Learn. Res.
**2014**, 15, 1929–1958. [Google Scholar] - Ding, X.; Zhang, X.; Ma, N.; Han, J.; Ding, G.; Sun, J. Repvgg: Making vgg-style convnets great again. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Nashville, TN, USA, 20–25 June 2021; pp. 13733–13742. [Google Scholar]

**Figure 1.**Examples of leaves/fruit of citrus pests and diseases. (

**a**) Citrus canker disease. (

**b**) Citrus scab disease. (

**c**) Citrus leaf miner insect-pests. (

**d**) Rust wall insect-pests. (

**e**) Citrus normal leaf/fruit picture.

**Figure 3.**Performance analysis of different transfer learning network models. (

**a**) F1 score of different OplusVNet models on various citrus pests and diseases. (

**b**) The accuracy of different OplusVNet models on the validation set.

Canker | Scab | Leaf Miner | Rust Wall | Normal | |
---|---|---|---|---|---|

Total | 1040 | 293 | 320 | 210 | 202 |

Training set | 624 | 176 | 192 | 126 | 122 |

Validation set | 208 | 59 | 64 | 42 | 40 |

Test set | 208 | 58 | 64 | 42 | 40 |

**Table 2.**The structure and parameters of the VGG16 transfer learning network of OplusVNet_10 with the input data size of $512\times 512\times 3$.

Number of Layers | VGG16 | Output Shape | Number of Parameters | |
---|---|---|---|---|

Frozen layer | 1 | $Conv\_1$ | $512\times 512\times 64$ | 1792 |

2 | $Conv\_2$ | $512\times 512\times 64$ | 36,928 | |

3 | $MaxPooling\_1$ | $256\times 256\times 64$ | 0 | |

4 | $Conv\_3$ | $256\times 256\times 128$ | 73,856 | |

5 | $Conv\_4$ | $256\times 256\times 128$ | 147,584 | |

6 | $MaxPooling\_2$ | $128\times 128\times 128$ | 0 | |

7 | $Conv\_5$ | $128\times 128\times 256$ | 295,168 | |

8 | $Conv\_6$ | $128\times 128\times 256$ | 590,080 | |

9 | $Conv\_7$ | $128\times 128\times 256$ | 590,080 | |

10 | $MaxPooling\_3$ | $64\times 64\times 256$ | 0 | |

Non-freezing layer | 11 | $Conv\_8$ | $64\times 64\times 512$ | 1,180,160 |

12 | $Conv\_9$ | $64\times 64\times 512$ | 2,359,808 | |

13 | $Conv\_10$ | $64\times 64\times 512$ | 2,359,808 | |

14 | $MaxPooling\_4$ | $32\times 32\times 512$ | 0 | |

15 | $Conv\_11$ | $32\times 32\times 512$ | 2,359,808 | |

16 | $Conv\_12$ | $32\times 32\times 512$ | 2,359,808 | |

17 | $Conv\_13$ | $32\times 32\times 512$ | 2,359,808 | |

18 | $MaxPooling\_5$ | $16\times 16\times 512$ | 0 |

**Table 3.**The parameters and structure of our proposed network with the input data size of $16\times 16\times 512$.

Layers | Layer Type | Core Size/Number | Convolution Step | Output Shape | Parameters |
---|---|---|---|---|---|

1 | $Conv\_1$ | $3\times 3/64$ | $1\times 1$ | $16\times 16\times 64$ | 294,976 |

$PReLU\_1$ | 64 | ||||

2 | $Conv\_2$ | $3\times 3/64$ | $1\times 1$ | 36,928 | |

$PReLU\_2$ | 64 | ||||

3 | $MaxPooling\_1$ | $3\times 3$ | $2\times 2$ | $8\times 8\times 64$ | 0 |

4 | $Conv\_3$ | $3\times 3/128$ | $1\times 1$ | $8\times 8\times 128$ | 73,856 |

$PReLU\_3$ | 128 | ||||

5 | $Conv\_4$ | $3\times 3/128$ | $1\times 1$ | 147,584 | |

$PReLU\_5$ | 128 | ||||

6 | $MaxPooling\_2$ | $3\times 3$ | $2\times 2$ | $4\times 4\times 128$ | 0 |

7 | $Conv\_5$ | $3\times 3/256$ | $1\times 1$ | $4\times 4\times 256$ | 259,168 |

$PReLU\_5$ | 256 | ||||

8 | $MaxPooling\_3$ | $3\times 3$ | $2\times 2$ | $2\times 2\times 256$ | 0 |

9 | $Conv\_6$ | $3\times 3/256$ | $1\times 1$ | 590,080 | |

$PReLU\_6$ | 256 | ||||

10 | $MaxPooling\_4$ | $3\times 3$ | $2\times 2$ | $1\times 1\times 256$ | 0 |

11 | $Flattern$ | 256 | 0 | ||

12 | $Fullyconnected\_1$ | $/512$ | 512 | 131,584 | |

$PReLU\_7$ | 256 | ||||

13 | $Fullyconnected\_2$ | $/5$ | 5 | 2565 | |

Total number of parameters | 1,538,149 |

Methods | F1 Score | Accuracy | ||||
---|---|---|---|---|---|---|

Canker | Scab | Leaf Miner | Rust Wall | Normal | ||

AlexNet | 0.98 | 0.88 | 0.93 | 0.87 | 0.83 | 0.93 |

TL-VGG16 | 0.94 | 0.91 | 0.74 | 0.85 | 0.68 | 0.88 |

RepVGG | 0.99 | 0.97 | 0.96 | 0.94 | 0.93 | 0.97 |

OplusVNet$\_10$ | 1.00 | 0.97 | 0.99 | 0.99 | 0.95 | 0.99 |

**Table 5.**The accuracy of the proposed OplusVNet with different frozen layer networks on different number of image training sets.

Number | Number of Frozen Layers | ||||||
---|---|---|---|---|---|---|---|

4 | 6 | 8 | 10 | 12 | 14 | 16 | |

170 | 0.97 | 0.97 | 0.97 | 0.97 | 0.96 | 0.94 | 0.91 |

120 | 0.95 | 0.96 | 0.94 | 0.95 | 0.96 | 0.91 | 0.89 |

70 | 0.93 | 0.94 | 0.94 | 0.92 | 0.93 | 0.89 | 0.89 |

Number | Network | F1 Score | Accuracy | ||||
---|---|---|---|---|---|---|---|

Canker | Scab | Leaf Miner | Rust Wall | Normal | |||

170 | AlexNet | 0.87 | 0.75 | 0.85 | 0.91 | 0.74 | 0.83 |

TL-VGG16 | 0.92 | 0.90 | 0.87 | 0.85 | 0.83 | 0.87 | |

OplusVNet_10 | 0.98 | 0.97 | 0.97 | 0.97 | 0.98 | 0.97 | |

OplusVNet_6 | 0.98 | 0.97 | 0.97 | 0.97 | 0.98 | 0.97 | |

120 | AlexNet | 0.81 | 0.75 | 0.85 | 0.87 | 0.70 | 0.80 |

TL-VGG16 | 0.90 | 0.90 | 0.84 | 0.88 | 0.75 | 0.85 | |

OplusVNet_10 | 0.93 | 0.94 | 0.97 | 0.96 | 0.96 | 0.95 | |

OplusVNet_6 | 0.94 | 0.95 | 0.98 | 0.97 | 0.97 | 0.96 | |

70 | AlexNet | 0.81 | 0.63 | 0.84 | 0.84 | 0.62 | 0.75 |

TL-VGG16 | 0.86 | 0.82 | 0.79 | 0.79 | 0.64 | 0.80 | |

OplusVNet_10 | 0.93 | 0.90 | 0.94 | 0.90 | 0.93 | 0.92 | |

OplusVNet_6 | 0.95 | 0.92 | 0.97 | 0.91 | 0.95 | 0.94 |

Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. |

© 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Yang, C.; Teng, Z.; Dong, C.; Lin, Y.; Chen, R.; Wang, J.
In-Field Citrus Disease Classification via Convolutional Neural Network from Smartphone Images. *Agriculture* **2022**, *12*, 1487.
https://doi.org/10.3390/agriculture12091487

**AMA Style**

Yang C, Teng Z, Dong C, Lin Y, Chen R, Wang J.
In-Field Citrus Disease Classification via Convolutional Neural Network from Smartphone Images. *Agriculture*. 2022; 12(9):1487.
https://doi.org/10.3390/agriculture12091487

**Chicago/Turabian Style**

Yang, Changcai, Zixuan Teng, Caixia Dong, Yaohai Lin, Riqing Chen, and Jian Wang.
2022. "In-Field Citrus Disease Classification via Convolutional Neural Network from Smartphone Images" *Agriculture* 12, no. 9: 1487.
https://doi.org/10.3390/agriculture12091487