Deep Learning-Based Mango Leaf Detection by Pre-Processing and Segmentation Techniques

Abstract

Different diseases afflict mango trees, and it can be quite challenging to see sickness with the naked eye.In contrast to the conventional system, the pre-processing and segmentation methods presented in this research for mango leaf disease detection use deep neural networks (DeNeuNet) to classify the segmented diseased part and make disease identification easier and more accurate.The goal of this project is to more effectively detect plant disease indicators using machine learning than a manual monitoring system. By using a classification technique to gather pictures of leaves that have various diseases affecting them, trained data are obtained in this case.A machine learning system is created to automatically upload and compare new photos of afflicted leaves with learned data in order to identify the symptoms of mangoes leaf diseases. Experimental results obtains by proposed technique is accuracy of 95.35%, precision of 90%, recall 88%, F-1 score 85%.