Hello, on assignment again. This time developing metrics measuring a project's impact of teaching better mining practices (i.e. "don't use mercury, it will kill trees") on the rain forest in Northwestern DRC, among other performance indicators. I am using a LANDSAT 8 'finished product': Surface Reflectance-Derived Spectral Index (NDVI) for 2011. There are many Saturate Values, which are out of the normal range of the NDVI Index (10k to -10k, for the LANDSAT 8 product). If I understand the saturate value correctly, I have two options. 1) 'recode' them out of the analysis (-9999) or 2) code them to the max, 10k value? Just like to mention that I have relatively short amount of time to conduct this analysis. I have other dimensions to measure such as the project's impact on miners' livelihoods (i.e. gender and economic analysis). Thanks so much.
in this case you can just replace them with -9999 or some other arbitrary value and cover the other impacts of project. There are no of reasons you are getting this outliers, may be you are not using accurate method, may be you do not understand NDVI formula. Use erdas beacuse it has build in model to run NDVI
- Badges
- Science topic
- Similar topics
- Geoscience
- Remote Sensing