International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 11 Issue: 06 | Jun 2024
www.irjet.net
p-ISSN: 2395-0072
Deep Learning-Based Rice Disease Recognition Using VGG16 and Transfer Learning Poreddy Jayaraju1, Dr. Aashiq Banu2 1 Phd Scholar ,Department of Computer science & Eng. Indian Institute of Technology Bhilai,
Chhattisgarh, India.
2Assistant Professor Department of Computer Science Eng. Koneru Lakshmaiah Education Foundation,
Hyderabad-500075, Telangana, India. ---------------------------------------------------------------------***--------------------------------------------------------------------networks used most frequently in deep learning, Amit Abstract - Rice is one of the major cultivated crops in India
Kumar Singh et al. used Support Vector Machine (SVM) to classify normal rice leaves and diseased rice leaves, and the classification accuracy reached 82% (Duan et al., 2017). Mohsen Azadbakht et al. used wheat leaf hyperspectral data and machine learning methods to detect wheat leaf rust, and reached a conclusion that support vector regression had the best effects in this case after comparing the results of four machine learning methods (Azadbakht et al., 2019). Comparatively speaking; the performance is very ordinary. Unlike traditional machine vision algorithms, which require manual feature extraction and classification, CNN only needs to input the image data into the network, and the selflearning ability of the network can complete the image classification (Xie et al., 2020), this research uses CNN as the research method.
which is affected by various diseases at various stages of its cultivation. It is very difficult for the farmers to manually identify these diseases accurately with their limited knowledge. For disease management, farmers spending lot of time and resources and they detect the diseases through their penniless naked eye approach which leads to unhealthy farming. The advancement of technical support in agriculture greatly assists for automatic identification of infectious organisms in the rice plants leaves. The convolutional neural network algorithm (CNN) is one of the algorithms in deep learning has been triumphantly invoked for solving computer vision problems like image classification, object segmentation, image analysis, etc. In our work, VGG16 is a type of CNN model utilized with transfer learning approach for recognizing diseases in rice leaf images. The parameters of the proposed model is optimized for the classification task and obtained a good accuracy of 97.57%.
Deep learning has been applied rapidly in image recognition (Xiong et al., 2021; Naranjo-Torres et al., Krishnamoorthy N et al. (2021) proposed a InceptionResNetV2 (Convolutional Neural Network) with transfer learning to identify rice plant diseases (Krishnamoorthy N et al. 2021). In this classification Solemane Coulibalya et al. (2019) applied a VGG16 model with transfer learning approach for detecting the disease in millet crop (Coulibaly et al., 2019). This work collected 124 leaf images and split into mildew diseases and healthy categories. The VGG16 model obtained an accuracy of 95%
Key Words: CNN Deep Learning Fine-tuning Rice leaf diseases Transfer learning
1.INTRODUCTION Timely and accurate diagnosis of plant diseases is crucial for sustainable agriculture and resource management. While some diseases lack visible symptoms, most are identified through optical observation by experienced plant pathologists. However, climate changes and the spread of diseases can challenge even skilled pathologists. In India, a major rice producer, agriculture contributes 19.9% to the GDP, with rice being a staple crop. Diseases affecting rice can significantly impact farmers' profits. Thus, an automatic data processing expert system for early disease detection is essential. Deep learning, particularly convolutional neural networks (CNNs), offers robust solutions for plant disease classification and other agricultural challenges by effectively processing visual data and learning spatial relationships.
N. Nandhini et al. (2020) proposed a machine learning algorithms such as SVM, K-NN, and decision trees for classifying the diseases in the plant leaves (Nandhini and Bhavani, 2020). Task, the authors used 6 classes of images for training and achieved an accuracy of 95.67% for InceptionResNetV2.
3. Materials and methods: In this section, the procedure of the projected work is segregated into seven steps processes for categorizing rice leaf diseases which shows in Fig. 1 (Fig. 2).
2.Related work A lot of researchers have developed disparate architectures in the recent years for the plant leaf diseases diagnosis by using machine learning and deep learning algorithms. 2020). Convolutional Neural Network (CNN) is one of neural
© 2024, IRJET
|
Impact Factor value: 8.226
3.1. Rice disease types and dataset description: The rice image dataset has been collected over the past few months mostly from the cultivation fields of Raipur,
|
ISO 9001:2008 Certified Journal
|
Page 192