Multigrids Image Analysis

Multigrids menu options on Image analysis Menu.  You must have loaded the satellite image into memory before this will work.

• Change display band
• Basic statistic: max, min, mean, std dev
• Scattergram of two bands
• Filter image
• Supervised classification
• Unsupervised classification
• Variance-covariance/Principal components
• Average reflectance in region
• Reinterpolate all to same resolution
  • Will use the resolution of the grid with the smallest data spacing
  • Will use bilinear interpolation
  • Will place files in a new subdirectory named "resolution_NN"
  • Be judicious and select only those bands you want to use; with Sentinel-2, this will increase the storage required by 3 times.
• Create Max DN grid:
• Create Average DN grid: create and open a grid with the average DN in all bands.
• Create Min DN grid
• Cloud brightening
  • Need a database with the shadowing percentage by band.  You could create this with two Average reflectance in region, and then do a percentage between the two.  This must be in the same directory as the image, and named "cloud_shading.dbf";
  • Need a grid with a mask for the clouds.  Outline a shapefile with the clouds, plot it on a grid, and mask out everything not colored.  This must be in the same directory as the image, and named "cloud_mask.dem";
  • Creates a new set of grids, one per band, in a new subdirectory.
• Locate values; create a database with all points within a given range.  This will have the LAT/LONG, BAND, and DN.  This database can become very large if you do not pick values judiciously.  It can be useful to find extreme values.
• Tail percentiles: create a database with key information on each band
• Normalize

Last revision 5/11/20187