Optimized Image Enhancement Enabled LSTM Technique for Plant Leaf Disease Detection

Diagnosing plant diseases is one of the most crucial and essential tasks in agriculture. Visual inspection or microscope is frequently used to identify causes. Visual judgment is subjective, and mistakes, biases, and optical illusions can arise due to the effect of cognitive and psychological processes and the subjective character of the job. Molecular biology, immunology, and pathogen culture-based laboratory studies sometimes take too long and do not produce results quickly enough. In light of this, automated technology's ability to quickly and accurately diagnose illness is essential. For big farms, manual observation is impracticable and costly since it requires constant plant monitoring. This article presents an optimized image enhancement enabled LSTM technique for plant leaf disease detection. Leaf images data set is used as input. The noise in these leaf images is Gotten rid of by employing an adaptive filter during the preprocessing stage of the development process. The Particle Swarm methodology is used by the LSTM optimization method in order to get a finer level of control over the input parameters. Additionally, Grey Wolf Optimization is used in order to enhance the functionality of the input parameters of the LSTM classifier. The approach known as the Gradient Boosting Decision Tree is further used for classification purposes. For the purpose of the experiment, the PlantVillage data set is used. Total 2500 images are selected at random for the experimental work. Two hundred photographs are used in order to train the classification model, and the remaining five hundred images are used in order to test the classification model after it has been trained. In the process of classifying leaf images, the PSO LSTM algorithm is becoming more accurate. PSO LSTM has achieved accuracy of 99 percent.