Deep Learning-based Disease Detection using Pomegranate Leaf Image
Mahesh Nirmal, Pramod Jadhav, Santosh Pawar, Manoj Kharde, Pravara
Abstract
The goal of this research is to detect a pomegranate plant leaf disease that will identify the diseases by making use of a deep convolutional neural network. Plant diseases are a serious problem in India and other Asian Countries that rely heavily on agriculture. Throughout the course of the year, several diseases can be found causing havoc on the harvest by attacking crops. Plant diseases can be difficult to identify with the naked eye alone. As a consequence of this, the development of a system that is capable of recognizing diseases is of the utmost importance. This paper proposes a deep learning technique to an image of a plant leaf, the disease detection model that has been suggested makes use of a deep convolutional neural network to locate and identify the disease. 447, 56, 56 pictures representing 14 unique species and 26 distinct diseases were utilized throughout the training process of the model. A CNN + LSTM is further developed with the help of a trained model. This proposed technique not only diagnoses a health problem, but it also suggests courses of treatment based on the information that it has gathered. In the vast majority of cases, farmers and other specialists in the sector keep a close eye on plants in order to detect and identify diseases. The proposed framework was developed with the assistance of deep learning technique. According to the findings of the tests, the framework that has been proposed is accurate to the degree of 90.546percent when it comes to differentiating between good and unhealthy leaves. The framework allows for the classification of diseases that affect pomegranate leaf to an accuracy of 97.246 %. The data sets are from Mendeley Data Total: 559 images. In which healthy 287 images were identified and 272 diseases images were identified. Originally data were split in 8:1:1 ratio.