Automated disease detection in crops using CNN
- 1. Automated Disease Diagnosis in Crops Seminar Guide Dr. Pradeep N Ph.d., Professor, Department of CS&E, BIET, Davangere VISVESVARAYA TECHNOLOGICAL UNIVERSITY BAPUJI INSTITUTE OF ENGINEERING TECHNOLOGY, DAVANGERE Department of Computer Science and Engineering TECHNICAL SEMINAR-2024 ON Technical Seminar Co-ordinators Prof. Rahima and Prof. Madhu Hiremath Asst. Professors, Department of CS&E, BIET, Davangere. -Manoj Bhavihal(4BD20CS051) Head of Department Dr. Nirmala C R Ph.d., Professor and Head, Department of CS&E, BIET, Davangere.
- 2. Contents • Abstract • Introduction • Objectives • Methodology • Implementation • Screenshots • Future Scope • Conclusion
- 3. Abstract Plant diseases affect the growth of their respective species, therefore their early identification is very important. Automatic leaf disease detection and classification systems allow farmers to better detect and manage crop diseases, potentially increasing productivity and farm sustainability. Agriculture field has a high impact on our life. Agriculture is the most important sector of our economy. Proper management leads to a profit in agricultural products. Farmers do not have expertise in leaf disease so they produce less production. Plant leaf diseases detection is important because profit and loss depend on production. CNN is the solution for leaf disease detection and classification. The main aim of this research is to detect the plants leaf diseases. Plant leaf disease detection has a wide range of applications available in various fields such as Biological Research and in Agriculture Institute. Plant leaf disease detection is one of the required research topics as it may prove beneficial in monitoring large fields of crops, and thus automatically detect the symptoms of diseases as soon as they appear on plant leaves.
- 4. Introduction Identification and classification of leaf diseases in agriculture are crucial for crop health. Traditionally, this task has been manual and error-prone. Recent advances in computer vision and machine learning, specifically Convolutional Neural Networks (CNNs), have enabled automated leaf disease detection and classification. Our approach utilizes image data of various diseases (e.g., powdery mildew, rust) collected from different crops. Types of Diseases: Healthy Leaves: Vibrant green coloration, smooth texture, and uniform appearance. Powdery Leaves: Characterized by a powdery white or gray substance on the surface. Rusty Leaves: Exhibit rusty or reddish-brown lesions, often accompanied by yellowing. CNN Architecture: Convolutional Layers : Extract features like edges, textures, or shapes. Pooling Layers : Reduce spatial dimensions while retaining important information. Activation Functions : Introduce non-linearity for learning complex relationships. Fully Connected Layers : Process high-level features and make predictions. Output Layer : Produces final predictions, such as healthy or diseased. Training Process: CNNs learn patterns and features relevant to leaf disease detection by adjusting network parameters using labeled training data. Prediction: Once trained, CNNs can make predictions on new data by propagating it through the network and selecting the class with the highest probability.
- 5. Objectives • To develop an automated system for the identification and classification of leaf diseases in agriculture. • To leverage recent advances in computer vision and deep learning to enhance the efficiency and accuracy of leaf disease detection. • To collect image data of various leaf diseases, including powdery and rusty infections, from different crops. • To design and implement a robust classification model capable of accurately distinguishing between healthy leaves and leaves affected by powdery or rusty infections. • To evaluate the performance of the proposed method and compare it with conventional approaches, demonstrating its superiority in terms of accuracy and effectiveness in detecting and isolating different types of leaf diseases..
- 6. Methodology 1. Data Acquisition : Collecting images of diseased and healthy leaves from various sources. 2. Preprocessing : Standardizing the images by resizing, cropping, and normalizing them. 3. Splitting Data : Dividing the dataset into training and testing sets to train and evaluate the model. 4. CNN Deep Learning Model: Developing a Convolutional Neural Network architecture for leaf disease detection. 5. Predictions : Using the trained model to make predictions on new, unseen leaf images.
- 7. Screenshots
- 11. Advantages 1. Automatic Feature Learning: CNNs can automatically learn relevant features from raw image data without the need for manual feature extraction. This ability to learn hierarchical representations of features can lead to more robust and accurate disease detection models. 2. Non-linear Decision Boundaries: CNNs are capable of learning complex non-linear decision boundaries, allowing them to capture intricate patterns and variations in leaf images that may not be captured by linear classifiers like SVM. 3. Scale-Invariant Representations: CNNs can learn scale-invariant representations of features, meaning they can detect diseases across leaves of different sizes and shapes without the need for explicit feature scaling or normalization. 4. Higher Accuracy: CNNs have demonstrated superior performance in leaf disease detection tasks compared to kNN and SVM, achieving higher accuracies and better generalization to unseen data.
- 12. Model SVM KNN CNN Accuracy 76% 64% 96%
- 13. Future Scope 1. Real-time Disease Detection on Mobile Devices : Developing lightweight and efficient CNN models that can run on smartphones and other mobile devices. This allows farmers to analyze diseases in their crops right in the field. 2. Integration with Drones and Robotics : Using CNN-based disease detection along with drones for large-scale monitoring of fields. Robots can then be deployed to intervene in affected areas autonomously. 3. Explainable AI for Disease Diagnosis : While CNNs are powerful, their decision-making process can be hard to understand. In the future, we can use Explainable AI (XAI) techniques to explain why the CNN identifies certain diseases. This helps build trust in the system and provides insights into the features triggering disease identification.
- 14. Conclusion Automatic leaf disease detection and classification systems offer significant potential to revolutionize agriculture by enhancing crop disease control and mitigating their economic and environmental consequences. Convolutional neural networks have emerged as valuable tools for detecting and classifying various leaf diseases across different plants and regions. While these machine learning models have demonstrated remarkable success, their performance can be influenced by the availability and quality of training data. Therefore, future research should prioritize the development of methods to acquire and enhance training data to enhance the accuracy and efficacy of these models. In conclusion, ongoing research in leaf disease detection and classification is crucial for the sustainability and advancement of agriculture. By refining accurate and efficient methods, we can empower farmers to better manage crop diseases, thereby minimizing their economic and environmental impact and ensuring food security for the future.
- 15. Thank You!