Hibiscus Plant Leaf Disease Detection using Modified Sigmoid Function in Logistic Regression

Abstract

Hibiscus plants are popular improvements known for their vibrant blooms and lush foliage. However, like all plants, they are disposed to various diseases that can poorly affect their health and aesthetics. As tropical plants, hibiscus plants need full sun to limited shade to thrive. Too much direct sunlight can result in leaf sunburn, causing little white spots to appear on the foliage. Early detection of these diseases is critical for timely participation and active management. In recent years, progressions in image processing and machine learning have offered promising solutions for mechanized disease detection in plants. This study proposes a novel approach for the automated detection of hibiscus plant leaf diseases using machine learning techniques. Digital images of hibiscus leaves are developed using a high-resolution camera or smartphone camera. Early detection and classification of diseases in hibiscus plants are dangerous for effective plant management and disease control. To evaluate the performance of proposed method, a dataset of labelled hibiscus leaf images containing different disease types. The extracted features are used to train a machine learning model to classify the images into healthy or diseased categories. For identification and classification of disease type, the existing method is SVM Classifier with “Modified RBF kernel” and the proposed method is “Logistic Regression with Modified Hyperbolic Tangent Function”. The method's efficiency can be assessed by employing metrics such as accuracy, sensitivity, specificity, and F1 score. Future research directions include further optimization of the classification model and integration of additional image processing techniques aimed at improved performance in real-world applications.