Design and Development of Classification Evaluation Framework for Apple Leaves Diseases Using Machine Learning Models

Abstract

Apple is one of the utmost consumed fruits in the globe and has a considerable impact pr newlinesignificance in the agriculture sector. However, the Fungal diseases especially those affecting newlineleaves, are a common and serious problem in apple orchards. These diseases not only worsen newlinethe quality of the leaves but also have a knock-on effect on the complete health and yield of newlinethe apple tree. Therefore, accurate diagnosis and localization of plant diseases is significant to newlinemaximize fruit yield, minimize economic losses, ensure good quality and food security. The newlinetraditional disease detection methods often associated with several challenges and fails to newlineprovide timely diagnosis of plant leaf disease. In addition, it involves collecting samples (often newlinesubjected to human error), sending them to laboratories for analysis, and waiting for results, newlinecausing delay that may allow diseases to spread, impact yield and increasing control costs. newlineTraditional image processing methods often struggle to distinguish between similar disease newlinesymptoms or detect early-stage infections with minor visual signs. This can lead to newlinemisdiagnosis and missed opportunities for timely treatment. Weather conditions, light newlinevariations, and overlapping symptoms can complicate accurate visual assessments, and may newlineresults in sub-optimal management decisions. These limitations highlight the urgent need for newlineautomated, and efficient plant disease identification and classification solutions. newlineThe research work in this thesis introduces a novel end-to-end framework th newline