ANALYSIS OF AIRBORNE HYPERSPECTRAL IMAGE USING VEGETATION INDICES, RED EDGE POSITION AND CONTINUUM REMOVAL FOR DETECTION OF Ganoderma DISEASE IN OIL PALM
ANALYSIS OF AIRBORNE HYPERSPECTRAL IMAGE USING VEGETATION INDICES, RED EDGE POSITION AND CONTINUUM REMOVAL FOR DETECTION OF Ganoderma DISEASE IN OIL PALM
复制标题
DOI：
10.21894/jopr.2018.0037
复制
发表时间：
2018-09-01
影响因子：
1.3
通讯作者：
Boyd, D.
中科院分区：
文献类型：
--
作者：
Izzuddin, M. A.;Nisfariza, M. N.;Boyd, D.
由灵芝引起的油棕茎基腐病给马来西亚油棕产业带来了巨大损失。航空高光谱遥感技术可以提供帮助，以检测和分类不同类别的灵芝疾病严重程度指数（DSI）的油棕。应用5种常用植被指数（VI）、4种红边位置指数（REP）和4种连续体去除指数（CR）对油棕灵芝病进行分类，将油棕灵芝病分为T1（健康）、T2（轻度）和T3（重度）3类。采用混淆矩阵和t检验评价VI、REP和CR的准确性。结果表明，简单比率指数（SRI）和增强植被指数（EVI）对油棕灵芝病害的检测能力中等。SRI的分类准确率为44.4%，而EVI的准确率为40.7%;而其他三种VI的准确率较差（
The basal stem rot (BSR) of oil palm caused by Ganoderma has brought huge losses to the oil palm industry in Malaysia. Airborne hyperspectral remote sensing technology may provide assistance to detect and classify different categories of Ganoderma disease severity index (DSI) in oil palm. In this study, five common vegetation indices (VI), four red edge position (REP) and four continuum removal (CR) were applied to categorise oil palm into T1 (healthy), T2 (mild) and T3 (severe) infection of Ganoderma disease in oil palm. The accuracy of the VI, REP and CR were assessed using confusion matrix and t-test. The results revealed that two VI, i.e. Simple Ratio Index (SRI) and Enhanced Vegetation Index (EVI) have moderate capability for the detection of Ganoderma disease in oil palm. SRI showed moderate classification accuracy (44.4%) compared to EVI with 40.7% accuracy; while the other three VI had poor accuracy (