Lettuce Leaf Disease Detection Using Convolutional Neural Network Algorithm
Plant disease is a major problem towards agriculture, as some of the disease could be infectious, the farmer who are not expert in observing plant disease may lead to the disaster of plant dying. Lettuce is a vegetable that is usually served as salad because of the taste crisp and mild. Although let...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Year Project Report |
| Language: | English English |
| Published: |
Universiti Malaysia Sarawak, (UNIMAS)
2023
|
| Subjects: | |
| Online Access: | http://ir.unimas.my/id/eprint/44263/1/Ngu%20Su%20Hang%2024pgs.pdf http://ir.unimas.my/id/eprint/44263/5/Ngu%20Su%20Hang%20ft.pdf http://ir.unimas.my/id/eprint/44263/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Plant disease is a major problem towards agriculture, as some of the disease could be infectious, the farmer who are not expert in observing plant disease may lead to the disaster of plant dying. Lettuce is a vegetable that is usually served as salad because of the taste crisp and mild. Although lettuce is a cool season crop, it can be grown in Malaysia by controlling the temperature and the environment. The examples of lettuce disease are Powdery Mildew, Downy Mildew, Bacterial Leaf Spot, and the infection of Mosaic Virus. The diseased lettuce can be healed if it is observed in early stage, but the lesion of disease area in early stage is hard to observe with raw eye. Therefore, this project proposed a lettuce leaf disease detection application using deep learning algorithm which is the convolutional neural network (CNN) to classify whether the image of the lettuce leaf is healthy or diseased. The detection algorithm will be develop based on a modified AlexNet model. The input dataset for the training of model is the images of healthy lettuce, bacterial leaf spot diseased lettuce and powdery mildew diseased lettuce. The images all are undergoing
image processing to enrich the image dataset, improve the performance of the model and avoid overfitting problem. Each image will be labelled with the class for the CNN model to classify it. The image dataset will split into three set, training, validation and the testing. The evaluation of the model will be looking at the performance metrics which are precision, recall, F1 score and accuracy. The trained CNN model will then implement using OpenCV for real-time operation and Python language for the programming. |
|---|