Efficient Plant Disease Prediction based on Convolutional Neural Network using Optimized Proposed Logistic Decision Regression

Abstract

Agriculture nature is important for growing plants with supports of artificial intelligence. This work aims to detect the disease in the leaves, realizing the image analysis and classification technology. Manual identification of medicinal plants is a time-consuming process that requires the help of plant identification experts and manual identification of medicinal plants is a time-consuming process that requires the help of plant identification experts. Specifically, there are several innovations in image segmentation and recognition system for plant disease detection. In this way, to proposed Logistic Decision Regression (LDR) algorithm and Convolutional Neural Network (CNN) is implemented detecting the feature selection and classification. Initially the preprocessing and filter process correction task is usually performed by the wrapping filters. Then LDR feature selection is used to select the best features of medicinal plants for reducing classification problems. Leaves are most used to identify medicinal plants, also stems, flowers, petals, seeds, and even the entire plant used in an automated process. An automated disease detection system is based on the development of changes in the disease status of the plant's leave. For Convolutional Neural Network (CNN), it uses a complex feed-forward neural network, and a CNN has high accuracy in image classification and recognition. After evaluating the results of different image training library systems, effective image recognition function has been demonstrated to have high precision and strong reliability.