Optimization of hyperspectral remote sensing imagery for detection of Ganoderma disease in oil palm


Citation

Izzuddin M. A., . and Idris A. S., . and Nisfariza M. N., . (2013) Optimization of hyperspectral remote sensing imagery for detection of Ganoderma disease in oil palm. [Proceedings Paper]

Abstract

Airborne hyperspectral imaging is an advance imaging system that provides narrow and contiguous imagery. The imagery can be utilised to noninvasively detect crop disease. In this study an advanced image processing and classification method is proposed to analyse hyperspectral image data for detection of Ganoderma disease in oil palm. Support Vector Machines SVM were conducted and evaluated for classifying hyperspectral data in order to enhance the detection of Ganoderma disease in oil palm. The method was conducted on original hyperspectral imagery and also the first derivative of original hyperspectral imagery. Tree crown detection method was also applied to the imageries to pin-point the coordinates of oil palms infected with Ganoderma disease. The results showed that SVM improved the classification accuracy compared to simple classification technique. In addition the tree crown detection method improved the coordinate identification of infected oil palms in plantation.


Download File

Full text available from:

Abstract

Airborne hyperspectral imaging is an advance imaging system that provides narrow and contiguous imagery. The imagery can be utilised to noninvasively detect crop disease. In this study an advanced image processing and classification method is proposed to analyse hyperspectral image data for detection of Ganoderma disease in oil palm. Support Vector Machines SVM were conducted and evaluated for classifying hyperspectral data in order to enhance the detection of Ganoderma disease in oil palm. The method was conducted on original hyperspectral imagery and also the first derivative of original hyperspectral imagery. Tree crown detection method was also applied to the imageries to pin-point the coordinates of oil palms infected with Ganoderma disease. The results showed that SVM improved the classification accuracy compared to simple classification technique. In addition the tree crown detection method improved the coordinate identification of infected oil palms in plantation.

Additional Metadata

[error in script]
Item Type: Proceedings Paper
Additional Information: Available at Perpustakaan Sultan Abdul Samad Universiti Putra Malaysia 43400 UPM Serdang Selangor Malaysia. SB608 O27M939 2013 Call Number.
AGROVOC Term: Oil palm
AGROVOC Term: Plant diseases
AGROVOC Term: Disease recognition
AGROVOC Term: Ganoderma
AGROVOC Term: Rots
AGROVOC Term: Remote sensing
AGROVOC Term: Satellite imagery
AGROVOC Term: Imagery
AGROVOC Term: Classification
AGROVOC Term: Infection
Geographical Term: MALAYSIA
Depositing User: Ms. Suzila Mohamad Kasim
Last Modified: 24 Apr 2025 05:15
URI: http://webagris.upm.edu.my/id/eprint/13215

Actions (login required)

View Item View Item