Abstract: In India, paddy is one of the most widely grown crops. These days, this crop is facing challenges from diseases that affect its quality and yield. This study presents an effective machine learning and image processing-based analysis and disease detection system for paddy crops. In this work, totally three different disease classes were taken those were brown spot, leaf smut, Bacterial Leaf Blight, and healthy class are taken. The proposed system uses transfer learning from a pre-trained VGG16 convolutional neural network and fine-tunes the model parameters through hyperparameter tuning via grid search to optimize the SVM classifier. By doing so, an accuracy of 98% on the test dataset is achieved. The system also uses image processing techniques, including color thresholding, morphology operations, and contour detection, to analyse and quantify the affected area of diseased leaves. Moreover, the system provides remediation guidance for each disease, utilizing text-to-speech synthesis for multilingual accessibility.
Keywords: VGG16, SVM, Grid Search, Hyperparameter tuning, OpenCV2, Scikit-learn, Morphology operations, Contours, Image processing, gTTs.
Cite:
Siva Parvathi V, Pavan Gopi Chand Pidikiti, Juber Shaik, Nandu Rettapalli, Jayadeep Mothukuri,"Efficient Analysis and Disease Detection System for Paddy Crop Using Machine Learning and Image Processing Techniques", IJARCCE International Journal of Advanced Research in Computer and Communication Engineering, vol. 13, no. 3, 2024, Crossref https://doi.org/10.17148/IJARCCE.2024.13390.
|
DOI:
10.17148/IJARCCE.2024.13390
