Please use this identifier to cite or link to this item: https://dl.ucsc.cmb.ac.lk/jspui/handle/123456789/4630
Title: Diagnosis of bacterial leaf blight, brown spots and leaf smut rice plant diseases.
Authors: Jayasooriya, G. R. I. L.
Issue Date: 9-Aug-2022
Abstract: Rice is the most widely consumed food product and one of the extensively cultivated crops in Sri Lanka. Considering the human population, food is one of the major problems that Sri Lanka might be facing in the near future. Therefore, increasing the crop yield is one of the major needs of the country. When rice crops are infected with diseases, it results in a loss of crops. Therefore it is important to identify the disease in the early stage of infection to prevent the damage that can be done. Disease identification is very difficult without having a clear understanding. With the advancement of new technologies, researchers are interested in identifying paddy diseases through machine learning and image processing techniques to help farmers to identify infectious diseases accurately. It is a challenge to make observations that are closer to how the human eye is capable of observing the paddy leaf to diagnose the infected disease. In this project, an algorithm was developed to check whether the image contains different changes to the paddy leaf by considering the green color pixels and their variance. OpenCV libraries have been used to develop the algorithm for feature extraction and those features were used as attributes to the LightGBM algorithm to classify the disease images with over 80% accuracy. Recent Deep Learning model developments have shown that automated image recognition systems with Convolutional Neural Network (CNN) models can be accurate in such problems. Since the Rice leaf disease image dataset is not available, the UCI machine learning repository provided dataset has been used to develop paddy leaf disease, classification models with data augmentation techniques to increase the number of images of the dataset while developing the model. OpenCV based algorithm has been developed as a python HTTP (Hypertext transfer protocol) service including the ability to identify healthy paddy leaf and leaf validation. A web-based program has been developed using the CNN model with an image processing algorithm to diagnose whether a paddy leaf is infected with bacterial leaf blight, leaf smut, or brown spot using an infected paddy leaf image. Image preprocessing and segmentation techniques have been used to increase the accuracy of the model. This developed system will be useful for people who do not have much knowledge about symptoms of paddy diseases.
URI: https://dl.ucsc.cmb.ac.lk/jspui/handle/123456789/4630
Appears in Collections:2021

Files in This Item:
File Description SizeFormat 
2018 MCS 041.pdf1.9 MBAdobe PDFView/Open


Items in UCSC Digital Library are protected by copyright, with all rights reserved, unless otherwise indicated.