Ganoderma Disease in Oil Palm Trees Using Hyperspectral Imaging and Machine Learning

The oil palm industry is one of the main contributors to the Malaysian national economy. However, the industry faces the challenge of Ganoderma infection, a destructive disease affecting oil palm trees that causes base stem rot (BSR). To address this issue, a novel method is proposed for detecting Ganoderma disease in oil palm trees by combining hyperspectral imaging and machine learning techniques. In this paper, four different classifiers, such as Support Vector Machine (SVM), Random Forest (RF), Multilayer Perceptron (MLP), and Decision Tree (DT), were trained on a tabular dataset generated by clustering spectral signatures of each pixel in the hyperspectral image. The evaluation results show that the MLP model showed 100% accuracy, sensitivity, and specificity in distinguishing between healthy and infected trees. SVM with radial basis function (RBF) and polynomial kernels show the second highest performance, attaining 91.67% accuracy, 83.33% sensitivity, and 100% specificity. Before training on classifiers, a feature-based band alignment technique is developed to overcome the hyperspectral image misalignment issue and ensure the accuracy of the spectral data. The developed band alignment significantly improved spectral coherence across the adjacent bands, as proved by enhanced Root Mean Squared Error (RMSE), Pearson Correlation Coefficient (PCC), and Normalized Mutual Information (NMI) values. Additionally, the developed background subtraction method effectively extracted the tree pixels from the background, resulting in a high precision score of 93.99%. Future research will focus on collecting more data, incorporating temporal information for disease stage classification, and implementing a more robust machine learning (ML) model for performance enhancement. Overall, this research shows the great potential of hyperspectral imaging for accurate detection of Ganodermadisease in oil palm plantations. Doi: 10.28991/HEF-2025-06-01-05 Full Text: PDF