Mid-infrared spectroscopy for early detection of basal stem rot disease in oil palm

• This study develops a mid-infrared spectroscopy technique that can be used for early and accurate detection of Ganoderma infection with different severities in oil palm.
• Application of this technique could provide more efficient and timely management of Gamoderma than the current ground scouting.
• Unlike previous work, this technique can detect the mildly infected trees, even at the asymptomatic stage.

Basal stem rot (BSR), caused by Ganoderma boninense is known as the most destructive disease in oil palm plantations in Southeast Asia. Ganoderma could reduce the productivity of oil palm plantations and potentially reduce the market value of palm oil in Malaysia. Early disease management of Ganoderma could prevent production losses and reduce the use of chemicals. This study focuses on the development of a statistical model for the discrimination of Ganoderma infestation on oil palm trees at different stages using a Fourier transform infrared (FT-IR) spectroscopic technique. Leaf samples of healthy, mild, moderately, and severely-infected trees were measured using FT-IR spectrometers to obtain absorbance data from the range of 2.55–25.05 μm s (3921–399 cm−1). The samples were analyzed with and without dilution with KBr. After pre-processing (baseline correction and normalization), the Savitzky–Golay method was used to calculate first and second derivatives. Then, for the preprocessed raw, first derivatives and second derivatives datasets, principal component analysis was performed to reduce the dimensionality of the data. The selected principal component scores were used in classification using linear discriminant analysis (LDA), quadratic discriminant analysis (QDA), k-nearest neighbor (kNN) and Naive-Bayes (NB) multivariate classification algorithms. The algorithms were tested to classify the leaf samples into four levels of disease severity. The results demonstrated that when samples were prepared with KBr, the LDA-based model resulted in the highest average overall classification accuracy of 92%, with individual classification accuracies greater than about 90% using the pre-processed raw dataset. This verifies the potential of mid-infrared spectroscopy for Ganoderma detection in early symptomless stages of infection in oil palm.