USING REMOTE SENSING INDICES TO REDUCE EFFECTS OF HILLSHADE ON LANDSAT 8 IMAGERY

Phung Van Khoa; Nguyen Hai Hoa; Do Anh Tuan

Keywords:

Hillshade effects, Landsat 8, minimisation, remote sensing indices

Abstract

This paper presents the effects of hillshade on the spectral values of Landsat and application of remote sensing image indices to reduce the influence of topography. This research uses the calibrated Landsat 8 images that are provided by the Center for Science and Earth Observation, known as the Earth Resources Observation and Science (EROS), the United States Geological Survey (USGS). The research has conducted in three typical areas of mountainous terrain in Vietnam and has used 22 remote sensing indices, including NDVI, NBR, NBR2, NDMI, SAVI, MSAVI, EVI, RATIO, TVI, CTVI, TTVI, DVI, IRVI, NLI, OSAVI, RDVI, VARI, MSR, LAI, GNDVI, GRVI and RVI to assess the levels of each hillshade- adjusted index. The results showed that in the same forest type, the values of spectral band image on the hillshade side are lower than those on the sunshine side. The more pronounced the hillshade is, the greater the difference is. In the case of light hillsade, indices of NDMI, NBR, NBR2 and VARI are more likely to minimize the most effects of mountain hillsade, followed by the group of indices, namely NDVI, RVI, GRVI, MSR, OSAVI, IPVI, RATIO, GNDVI, CTVI, TTVI and TVI. The remaining group of indices including MSAVI, SAVI, EVI, LAI, RDVI, NLI and DVI are strongly affected by the terrain even in the case of light hillsade. The findings of this research are useful for selecting the remote sensing indices in relation to management of natural resources and environment where there is an existence of complex mountainous terrain.