A fast algorithm for automatic removal of mirror side banding from MODIS oceancolor imagery
MODIS imagery has long been plagued by optical errors (i.e. banding) introduced by the mirrors on board the spacecraft with a potentially significant impact on scientific study. Banding errors are most evident in the blue channels (412 nm, 440 nm, 490 nm) which are essential for chlorophyll retrieval algorithms. The error parameterization is also influenced by target and environmental factors that make it difficult to fully quantify spatially. The proposed algorithm analyzes a scene using Fourier transformation and computes a spatially dynamic correction factor. Resulting error is typically less than one percent of the underlying signal, tested by artificially introducing banding into a SeaWiFS scene which does not suffer from the mirror side banding problem. Visual analysis also confirms removal of the banding. The algorithm is robust and fast, allowing for batch processing of many scenes without need for manual tuning.
IAGLR 57th Annual Conference on Great Lakes Research
Sawtell, R. W.,
Sayers, M. J.
A fast algorithm for automatic removal of mirror side banding from MODIS oceancolor imagery.
IAGLR 57th Annual Conference on Great Lakes Research.
Retrieved from: https://digitalcommons.mtu.edu/mtri_p/90