I'm analyzing MOD09A1 spectral data using GDAL/Python. The quality mask is a uint32 dataset and want to pull out the value of the first bit to mask the spectral data (1 == poor data; 0 == good data). When I read the HDF directly into a Numpy array without mosaicking the HDFs or applying gdalwarp to reproject/clip the matrices, and build my quality mask from the uint32 dataset (quality_mask = quality_data & 1), there are plenty of values in the matrix at which bit_1 == 0 and bit_1 == 1. IF, on the other hand, I process the HDF data with GDAL (gdalwarp to reproject, transform to GTiff, clip), suddenly ALL of the bit1 values are equal to zero, i.e. all of the data values are flagged as good quality, which is not the case. Any thoughts on how/why the uint32 data is getting transformed in this way as I run it through gdalwarp?
أكثر...
أكثر...