|
A New Method for Transforming Classifying Plant Diseases Using Wavelet
|
Ali Karimi   , Vahideh Naderifar |
|
|
|
Abstract: (725 Views) |
Rapid detection of plant diseases has always been an important challenge for the agricultural industry. One of the approaches that has been welcomed in this field is the use of image processing methods. The advantage of these methods is that they are automatic, fast, low cost, non-destructive and accurate. In this study, by processing the leaves of plants and agricultural products, while distinguishing healthy plants from unhealthy ones, their type of disease is automatically detected. To do this, deep learning methods based on several different architectures of Convolutional neural networks have been used. The proposed method in this research can be generalized to different plants and products as well as to several plants simultaneously. Designed networks are represented and evaluated using two different subsets of database dataset images. In this research, the proposed method algorithm was expressed. In general, a classification model was created from the patient's leaf by means of a wavelet transform to diagnose the type of plant and the disease of that leaf. In this dissertation, using transfer learning (transfer of learning in neural networks means a slight change of a deeply trained network to solve problem A so that without the need for re-training the whole network can be used to solve a completely different problem B) We compared the performance of ResNet50, GoogleNet, AlexNet networks and a deep network with a simple design and selected the best performance model among them for our work. Deep learning is a fast and evolving knowledge that has many implications for agricultural imaging. Machine learning algorithms, such as SVM backup vector devices, are often used for detection and classification. But they are often limited to the assumptions made when defining features.
|
Article number: 5 |
| Keywords: Wavelet, Algorithm, AlexNet, GoogleNet, ResNet50, Deep Learning |
|
|
Full-Text [PDF 918 kb]
(234 Downloads)
|
Type of Study: Research |
Subject:
General
Received: 2023/01/1 | Accepted: 2023/01/1 | Published: 2023/01/1
|
|
|
|
|
|
|
| Add your comments about this article |
|
|
|
