Abstract
This paper presents a novel approach for the early detection of potato plant diseases using deep learning techniques. The proposed method, AgriScanNet-18, is a multilayer convolutional neural network (CNN) that uses image-based analysis to identify various plant diseases. By training and evaluating the model on a potato leaf disease dataset, we achieved high accuracy of 99.30% for training and 99.28% for testing. Additionally, we developed a web app that facilitates the diagnosis of potato plant diseases by easily uploading images of leaves. In comparison with state-of-the-art models such as, VGG16, ResNet50, and VGG19, AgriScanNet-18 demonstrated improved identification accuracy of 8.66%, 3.61%, and 7.45%. In addition, Potato plant diseases can be managed and controlled using this technology to increase crop production and profitability.
Original language | English |
---|---|
Publication date | 2024 |
Number of pages | 18 |
DOIs | |
Publication status | Published - 2024 |
Event | Intelligent Systems Conference - Amsterdam, Netherlands Duration: 7 Sept 2023 → 8 Sept 2023 Conference number: 9 |
Conference
Conference | Intelligent Systems Conference |
---|---|
Number | 9 |
Country/Territory | Netherlands |
City | Amsterdam |
Period | 07/09/2023 → 08/09/2023 |
Keywords
- Computer Vision
- Deep Learning
- Potato Leaf Diseases
- Receiver Operating Characteristic
- Residual Neural Network
- Visual Geometry Group