Plant Leaf Disease Diagnostic System Built on Feature Extraction and Ensemble Classification

Abstract

Plants are a major source of energy for humans and animals alike. Plant leaves are the most common way for plants to communicate with the atmosphere. As a result, academics and academicians are responsible for investigating the problem and developing ways for recognizing disease-infected leaves. Farmers all across the world will be able to take immediate action to avoid their crops from getting severely damaged, so sparing the world from a potential economic crisis. Because manually detecting diseases may not be the ideal solution, a robotic methodology for detecting leaf ailments could benefit the agriculture industry while also enhancing crop output. The goal of this study is to improve classification outcomes by combining ensemble classification in conjunction with hybrid Law's mask, Gabor, SIFT, GLSM and LBP. The method proposed exemplifies the usage of ensemble classification. For comparison purpose Gabor and SIFT are utilized. The features employed are also vital in attaining the best results because ensemble classification has demonstrated to be more accurate. The experiments used sick leaf pictures of three plant types from the PlantVillage.