I am trying to run an NDVI classification of 2015 multispectral LT8 imagery in QGIS to assess the quality of vegetation cover in regards to habitat suitability for a primate species in northern Vietnam.
Can anyone tell me if I need to make a reflectance conversion of the bands that I will use (band 4 and 5) or can I just assign the bands and use the (NIR-RED)/(NIR+RED) equation through the raster calculator?
How do reflectance conversions such as DOS1 affect imagery analysis?
Best Answer
Using Dark Object Subtraction (DOS) can provide a a quick way to correct for atmospheric effects, which may be sufficient for preparing multispectral data for spectral indices such as NDVI. However, it will not be as accurate as using Landsat Surface Reflectance High Level Data Products for Landsat 8 that is generated from the L8SR algorithm. The L8SR is available through earthexplorer and you can download the data you need, and then calculate NDVI more accurately.
Since you are using Semi-Automatic Classification Plugin (SCP) for QGIS, there is an interesting research done to compare DOS1 Correction with USGS High Level Data Products of the Landsat 8 Surface Reflectance titled: Landsat 8 Surface Reflectance: a Comparison of DOS1 Correction and USGS High Level Data Products.
You can test the difference by comparing the NDVI calculated using raw data DN values of Landsat 8 before conversion to surface reflectance, using DOS1, and using L8SR High Level Data and see difference.