CITRUS DISEASES DETECTION & CLASSIFICATION USING DL AND ML MODELS: A SYSTEMATIC REVIEW

Abstract

Abstract - Citrus fruits, including lemons, mandarins, oranges, tangerines, grapefruits, and limes, are widely cultivated around the world. Citrus manufacturing enterprises generate a substantial amount of trash annually, with fifty percent of citrus peel being lost to various plant diseases. This paper provides a survey of the various methodologies applicable to the detection and classification of citrus plant leaf diseases. The page provides a comprehensive classification of citrus leaf diseases. Initially, the difficulties of each stage, which affect the accuracy of detection and classification, are described in depth. In addition, a comprehensive literature assessment of strategies for automated disease identification and classification is offered. In order to accomplish this, several picture preprocessing, segmentation, feature extraction, feature selection, and classification techniques are investigated. Discuss the significance of feature extraction and deep learning methods as well. The survey provides a detailed assessment of studies, analyzes their merits and weaknesses, and identifies more research concerns. According to the survey results, automated detection and classification approaches for citrus plant diseases are still in their infancy. To fully automate the detection and classification processes, therefore, new technologies are required. Comparative analysis of deep learning models currently used for citrus disease detection and classification. The creation of an updated model containing new characteristics and classifiers. Improvement of the proposed model's accuracy in the identification and classification of citrus illnesses. Citrus production and export have increased gradually over the past three decades, albeit at a slower rate than rival products such as mangoes, avocados, and melons. Citrus fruit production is severely affected by illnesses in its growing stages. The diseases develop not only on foliage but also on fruits. Hence, the presence of faults degrades the quality of fruits. The citrus fruits are evaluated in two ways, based first on the color of their skin and then on their size. So, it is necessary to assess citrus illnesses in order to prevent output losses. In addition, citrus fruit must be graded to facilitate its packaging in terms of its quality, so that the correct Citrus fruit values can be generated. This research examined and assessed various machine vision-based citrus disease prediction and postharvest citrus fruit grading approaches reported between 2010 and 2022. This study discusses the present successes, limits, and recommendations for future research on citrus illnesses and fruit grading.