Using a 2D Convolutional Neural Network (CNN) that uses prediction and classification techniques, we describe a novel method for automated tomato leaf disease detection in this work. Our dataset includes a wide range of classes, from healthy leaves to leaves with common diseases as bacterial spot, target spot, tomato leaf curl virus, blight, and septoria leaf spot. We carefully choose the dataset using methodical pre-processing and augmentation procedures to enable efficient training, validation, and testing. The CNN design combines convolutional and pooling layers with fully linked layers for accurate classification, effectively extracting significant characteristics from leaf images. By training the algorithm on this improved dataset, we are able to classify and predict tomato leaf diseases with amazing accuracy.
Introduction
Introduction:
Tomato leaf diseases threaten crop yield and food security. Traditional manual disease identification is time-consuming and error-prone. This study presents an automated, cost-effective method using 2D Convolutional Neural Networks (CNNs) to detect and classify tomato leaf diseases via image analysis. This deep learning-based solution enables early diagnosis, which is crucial for timely disease management.
Key Components:
1. Background on Plant Diseases:
Crop diseases, caused by pathogens like fungi, bacteria, and viruses, lead to major agricultural losses.
Visual inspection by experts is traditionally used but is inefficient.
CNNs have emerged as powerful tools in plant disease detection, improving accuracy and reducing subjectivity.
2. Deep Learning Overview:
Deep learning mimics the human brain through artificial neural networks.
It is especially impactful in computer vision, enabling models to learn features from image data autonomously.
3. Literature Review Highlights:
Several studies demonstrated the success of CNNs and transfer learning (e.g., ResNet, DenseNet, EfficientNet) for plant disease classification.
Limitations include dataset quality, background complexity, and the need for real-world conditions.
Approaches such as data augmentation, GAN-generated synthetic images, ensemble models, and federated learning were explored to address issues like class imbalance, privacy, and overfitting.
4. Methodology:
The proposed system consists of:
A. Data Loading:
A curated dataset with multiple tomato leaf classes (e.g., healthy, blight, bacterial spot, curl virus, septoria).
B. Data Preprocessing:
Techniques include resizing, normalization, and cleaning to improve data quality and prepare it for training.
C. Feature Extraction:
Relevant image features are selected using tools like PCA, enhancing the model’s ability to distinguish between disease types.
D. Model Training and Testing:
The 2D CNN model is trained and validated on the processed dataset using standard practices such as data splitting, optimization, and performance monitoring.
E. Prediction and Evaluation:
The trained model classifies new leaf images into disease categories.
Performance is evaluated using metrics like accuracy, precision, recall, and F1-score, often compared to expert-labeled ground truth.
5. Results and Performance:
The 2D CNN achieved high accuracy in identifying several tomato diseases and distinguishing them from healthy leaves.
The model proved robust in real-world scenarios thanks to effective preprocessing and training.
This system provides a reliable, scalable tool for automated plant disease detection, enabling early intervention and reduced crop losses.
Conclusion
In conclusion, the automated identification of tomato leaf diseases using a 2D Convolutional Neural Network (CNN) represents a significant advancement in agricultural technology. In addition to effective CNN model training and testing, meticulous feature selection, preprocessing, and dataset curation achieve remarkable accuracy in sickness prediction and categorization. This innovative method may assist lower crop losses and increase agricultural productivity by offering a practical choice for the early detection and management of tomato leaf diseases.
References
[1] Vasileiosbalafas, emmanouilkarantoumanis, Malamatilouta and Nikolaosploskas,Machine Learning and Deep Learning for Plant Disease Classification and Detection, IEEE Access ( Volume: 11), 16 October 2023 , 10.1109/ACCESS.2023.3324722 , https://ieeexplore.ieee.org/document/10286031.
[2] Diana susan joseph , (student member, ieee), Pranav m. pawar ,and Kaustubhchakradeo, Real-Time Plant Disease Dataset Development and Detection of Plant Disease Using Deep Learning, IEEE Access ( Volume: 12), 25 January 2024,10.1109/ACCESS.2024.3358333, https://ieeexplore.ieee.org/document/10414062
[3] Muhammad HammadSaleem, Johan Potgieter and Khalid MahmoodArif , Plant Disease Detection and Classification by Deep Learning, MDPI Plants 2019, 8(11), 468, https://doi.org/10.3390/plants8110468
[4] Emmanuel moupojou, Appolinairetagne, Florentretraint ,Anicettadonkemwa, Dongmowilfried, Hyppolitetapamo and Marcellinnkenlifack, FieldPlant: A Dataset of Field Plant Images for Plant Disease Detection and Classification With Deep Learning, IEEE Access ( Volume: 11), 29 March 2023
10.1109/ACCESS.2023.3263042, https://ieeexplore.ieee.org/document/10086516
[5] k. p. Asha rani and. Gowrishankar,Pathogen-Based Classification of Plant Diseases: A Deep Transfer Learning Approach for Intelligent Support Systems,IEEE Access ( Volume: 11), 09 June 2023,10.1109/ACCESS.2023.3284680, https://ieeexplore.ieee.org/document/10147120
[6] Mohammed SaeedAlzahrani and FawazWaselallahAlsaade, Transform and Deep Learning Algorithms for the Early Detection and Recognition of Tomato Leaf Disease, MDPI Agronomy 2023, 13(5), 1184, 22 April 2023,https://www.mdpi.com/2073-4395/13/5/1184
[7] Mahrintasfe , Akmnivrito , Fadi al machot ,Mohibullah , (member, ieee), and Habibullah, Deep Learning Based Models for Paddy Disease Identification and Classification: A Systematic Survey, IEEE Access ( Volume: 12), 27 June 2024,10.1109/ACCESS.2024.3419708, https://ieeexplore.ieee.org/abstract/document/10574294
[8] MeenakshiAggarwal , VikasKhullar , NitinGoyal , Rama Gautam , FahadAlblehai , MagdyElghatwary and Aman Singh, Federated Transfer Learning for Rice-Leaf Disease Classification across Multiclient Cross-Silo Datasets,MDPIAgronomy 2023, 13(10), 2483, 27 September 2023, https://doi.org/10.3390/agronomy13102483
[9] Zhichao Chen, Guoqiang Wang , Tao Lv and Xu Zhang, Using a Hybrid Convolutional Neural Network with a Transformer Model for Tomato Leaf Disease Detection, MDPI Agronomy 2024, 14(4), 673, 26 March 2024,https://doi.org/10.3390/agronomy14040673
[10] Weihao Su, Yutu Yang , Chenxin Zhou, ZilongZhuang and Ying Liu, Multiple Defect Classification Method for Green Plum Surfaces Based on Vision Transformer, MDPI Forests 2023, 14(7), 1323,28 June 2023, https://doi.org/10.3390/f14071323
[11] Jian P, Chen K, Cheng W (2021) GAN-Based One-Class Classification for Remote-Sensing Image Change Detection. IEEE Geosci Remote SensLetthttps://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=10560510
[12] Abdollahi A et al (2020) Building footprint extraction from high resolution aerial images using Generative Adversarial Network(GAN) architecture. IEEE Access 8:209517–209527https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9260150
[13] Le VNT et al (2020) A novel method for detecting morphologically similar crops and weeds based on the combination of contour masks and filtered Local Binary Pattern operators. GigaScience9(3): giaa017 https://academic.oup.com/gigascience/article/9/3/giaa017/5780256
[14] Albahli S et al (2021) An improved faster-RCNN model for handwritten character recognition. Arab J SciEng 1–15 https://www.researchgate.net/publication/350508860_An_improved_faster-RCNN_model_for_handwritten_character_recognition
[15] Albahli S et al (2021) Recognition and Detection of Diabetic Retinopathy Using Densenet-65 Based Faster-RCNN. ComputMaterContin 67:1333–1351 https://www.techscience.com/cmc/v67n2/41339
[16] Chen J, Yin H, Zhang D (2020) A self-adaptive classification method for plant disease detection using GMDH-Logistic model. Sustain Compute Inform Syst 28:100415 https://www.sciencedirect.com/science/article/abs/pii/S2210537920301426
[17] Atila Ü et al (2021) Plant leaf disease classification using efficient net deep learning model. Eco Inform 61:101182 https://www.sciencedirect.com/science/article/abs/pii/S1574954120301321
[18] Solano-Rojas B, Villalón-Fonseca R, Marín-Raventós G (2020) Alzheimer’s Disease Early Detection Using a Low Cost Three-dimensional Densenet-121 Architecture. In: International Conference on Smart Homes and Health Telematics. Springer https://link.springer.com/chapter/10.1007/978-3-030-51517-1_1
[19] Agarwal M et al (2020) ToLeD: Tomato leaf disease detection using convolution neural network. ProcediaComputSci 167:293–301 https://www.sciencedirect.com/science/article/pii/S1877050920306906
[20] Argüeso D et al (2020) Few-Shot Learning approach for plant disease classification using images taken in the field. Comput Electron Agric 175:105542 https://www.sciencedirect.com/science/article/abs/pii/S0168169920302544
[21] Singh, V. and Misra, A. K. (2021). \"Using image segmentation and soft computing techniques to detect leaf diseases.\" Agriculture information processing, 4, pp. 41–49.
[22] Lee and Chan (2020) Mayo, S.J., Lee, S.H., and Remagnino, P. \"Deep learning for plant classification: extracting and learning leaf features.\" p. 1–13 in Pattern Recognition, 71.
[23] Alchanatis, V., Hillel, A.B., Krikeb, O., and Farjon, G. (2021). \"Identifying and quantifying apple tree flowers to improve chemical thinning choices.\" 1–19 in Precision Agriculture, 21.
[24] Mohanty, S.P., Salathé, M., and Hughes, D.P. (2020). \"Detecting crop diseases using images and deep learning.\" Plant Science Frontiers, 7, p. 1419.
[25] Liu, W., Wu, G., and Ren, F. (2021). \"Reuse of residue networks for pest identification is proposed.\" Access IEEE, 7, pp. 122758–122768.
[26] A. Krizhevsky, I. Sutskever, and G. E. Hinton\'s paper \"ImageNet classification with deep convolutional neural networks\" was published in Proc. Adv. Neural Inf. Process. Syst. (NIPS), vol. 25, Dec. 2012, pp. 1097–1105.
[27] \"Going deeper with convolutions,\" by C. Szegedy, W. Liu, Y. Jia, P. Sermanet, S. Reed, D. Anguelov, D. Erhan, V. Vanhoucke, and A. Rabinovich, is cited in IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), June 2015, pp. 1–9.
[28] K. Simonyan and A. Zisserman, \"Very deep convolutional networks for large-scale image recognition,\" arXiv:1409.1556, 2014
[29] \"Deep residual learning for image recognition,\" Proceedings of the IEEE Conference on Computer Vision Pattern Recognition (CVPR), June 2016, pp. 770–778; K. He, X. Zhang, S. Ren, and J. Sun
[30] Convolutional networks with dense connections, K. Q. Weinberger, L. Van Der Maaten, Z. Liu, and G. Huang, Proceedings of the IEEE Conference on Computer Vision Pattern Recognition (CVPR), July 2017, pp. 4700–4708.