Detection of Basal Stem Rot (BSR) disease at oil palm plantation using hyperspectral imaging

Abstract

Basal Stem Rot (BSR) disease is major threat to oil palm plantation caused by fungi Ganoderma boninensis sp in which the symptom visually noticeable only at the late stage of infection. The objective of this study is to detect presence of BSR disease by using airborne hyperspectral imaging. The study site was located at Kertam Oil Palm Plantation, Sabah, Malaysia covering 5,000 ha area. Hyperspectral image was captured by flying at 1,000m altitude to achieve 1.0m × 1.0m spatial resolution image using the AISA Dual sensor. Upon samples collected, image was classified using Spectral Angular Mapper (SAM) classification. Verification was conducted by taking tissue samples from the basal stem of randomly selected palm at the site. Samples collected were subsequently tested in Ganoderma Selective Medium (GSM) in Petri dish to observe fungus presence. Results showed that from 51 palms suspected with BSR disease by hyperspectral imaging, 49 palms (96%) were found to be correct. Study results concluded that airborne hyperspectral imaging can be reliably applied to detect BSR disease even at early detection.