應用航拍高光譜影像與多光譜影像於農作判釋－以新竹為例

APPLYING AERIAL HYPERSPECTRAL IMAGING AND MULTI-SPECTRAL IMAGING TO AGRICULTURAL INTERPRETATION: A CASE STUDY OF HSIN-CHU AREA

林若琪(Jo-Chi Lin) ； 林保宏(Pao-Hung Lin) ； 方鋐亦(Hong Yi Fang) ； 林威延(Wei-Yen Lin)

《中國土木水利工程學刊》 30卷2期 (2018/06) Pp. 121-127

https://doi.org/10.6652/JoCICHE.201806_30(2).0004

摘要

本研究以航拍高光譜影像與航拍多光譜影像作為研究材料，針對新竹研究區域判釋農作物類別。航拍多光譜影像分為3波段與4波段2種組合，所有影像皆產製紋理與植生指標指數套疊原始影像，航拍高光譜則以MNF方法維度轉換後才進行紋理與植生指標套疊，結果顯示，航拍高光譜相比於航拍多光譜在相同波譜範圍間能產製較多光譜特徵影像，利於後續分類，紋理指標在各影像中，仍有其作用，未來應可針對不同尺度開發其相關影像使用潛能。

關鍵字

多光譜影像 ； 高光譜影像 ； 農作判釋

並列摘要

In this study, aerial hyperspectral images and aerial multi-spectral images are adopted as research materials and agricultural classification interpretation in the Hsin-chu research area is conducted. Aerial multi-spectral images are divided into 3-wave band and 4-wave band, in which all images generate texture nesting original images with vegetation indices. In aerial hyperspectral images, dimension conversion is conducted with the MNF method and then texture and vegetation index nesting is proceeding. The results show that compared to aerial multi-spectral images, aerial hyperspectral images can produce more images of spectral features in the same spectral range, which benefits follow-up classification. Texture indicator is still effective in each image. In the future, it is suggested to develop related imaging use potential for different scales.

並列關鍵字

multi-spectral imaging ； hyperspectral imaging ； agricultural interpretation