New answers tagged erdas-imagine
Use the Layer Stack tool in reverse by using your 11 band stack as the input and selecting only the bands you want in the output. For example, in the attached screenshot I added bands 1, 2, 3 to the stack (omitting band 4).
As mentioned by @nicholaschris, Zhu et al's paper is nice, and they have a tool associated with it. Note that the shift is a function of 1) the position of the sun (doesn't vary on one image), 2) the position (XYZ) of the cloud and 3) the elevation of the ground (2 and 3 do vary). So that a unique shift could not be enough. To answer the question with ...
You can also complete this analysis using ArcGIS or QGIS. If you are using indices derived from MODIS, just add them into the queue. Otherwise, you will need to separate the MODIS bands prior to adding them to the queue. In ArcGIS, use the Cell Statistics (Spatial Analyst) tool. In QGIS, use the r.series tool in the GRASS toolkit.
It turns out the problem was a lingering projection issue. The NDVI scenes were default projecting in Arc as a similar, but different projection (sorry, I can't remember what it was off the top of my head), whereas the assigned projection from USGS was Lambert Azimuth Equal Area. The projection along with a tutorial of how to correct the issues are posted at ...
Top 50 recent answers are included