Reconfigures ENVI imagery data (ascii file) for analysis using a data science programing language (e.g. Matlab or Python)
Sometimes ENVI users would prefer analyzing a masked image (or a image subset) with a programing language that is not IDL. For this purpose, ENVI allows user to download masked imagery data as ascii files. However these files are often exceedingly large; furthermore, their formatting is extremely confusing. This makes the task of organizing imagery data especially daunting. Fortunately, ENVI_MaskPixel_Harvester reconfigures ENVI-generated ascii data into easy-to-use Matlab structures/python dictionaries. It organizes imagery in two ways: 1. It isolates masked pixels in "regions of interest" (ROIs) allowing users easy access to reflectance (Rrs) spectra of all masked pixels while discarding pixels outside masks, 2. It compiles data spatially by band/wavelength (including both masked and unmasked regions of the original image).
Inputs: ENVI-generated ascii file, BIP interleaved
Outputs: MASK_pix - This structure contains the Rrs spectrum for each masked pixel ENVI_GRID - THis structure organizes all image pixel values (masked and unmasked) in two-dimensional matrices according to their band layers.
Required files: SENSOR_Bands.mat
Directions for use:
-
'Pick ASCII file' - User will be asked to select an ascii file for analysis. This file should be of the .txt file of an image interleaved as BIP
-
'Left Corner Pixel Coordinates' - In the event that the user cropped the Envi Standard File (file from which he/she downloaded the saved ascii file), matlab image coordinates will not correspond with the ENVI coordinates. This will compensate for the offset. The 'sample' (X) and 'line' coordinates for the upper left corner for an ENVI Standard File can be found in 'Edit Header' within ENVI Classic.
-
'Choose a sensor' - User can choose between AVIRIS, MODIS, Landsat TM (Landsat 4 & 5), Landsat 7, and Landsat 8. These terms are case sensitive.
-
'Enter band numbers' - User must enter the band numbers that his/her Envi Standard File (ESF) contains. For example, if the ascii file is saved from an AVIRIS ESF that containing bands 10-15 & 20, the user must input '10 11 12 13 14 15 20 '. Because these bands are consecutive, the user may also input '10:15 20'. The user may input these numbers out of order, as long as they are present.
Failure to do line up the a wavelenght with its proper band number this will create inaccurate pixel Rrs spectra. For each sensor, band numbers and their corresponding wavelengths (as read by MaskPixel_Harvester) are listed below.
If using landsat 4,5 or 7, user should crop out band 6 and bands 8+ in ENVI BEFORE exporting data in an ascii file. If using Landsat 8, user should crop out bands 8+ BEFORE exporting data in an ascii file. Corresponding sensor band numbers are listed below:
AVIRIS - bands 1-224 correspond to their registered wavelengths
MODIS (may differ from actual band numbers) Band 1 - 412nm Band 2 - 443nm Band 3 - 469nm Band 4 - 488nm Band 5 - 531nm Band 6 - 547nm Band 7 - 555nm Band 8 - 645nm Band 9 - 667nm Band 10- 678nm
Landsat TM (may differ from actual band numbers) Band 1 - 485nm Band 2 - 560nm Band 3 - 660nm Band 4 - 830nm Band 5 - 1650nm Band 6 - 2215nm (band 7 actual)
Landsat 7 (do not use Band 6 or 8) Band 1 - 477.5nm Band 2 - 560nm Band 3 - 661nm Band 4 - 835nm Band 5 - 1648nm Band 6 - 2204.5nm (band 7 actual)
Landsat 8 (do not use Band 8-11) Band 1 - 440nm Band 2 - 480nm Band 3 - 560nm Band 4 - 655nm Band 5 - 865nm Band 6 - 1610nm Band 7 - 2200nm
Outputs: MASK_pix.sample - the sample number (ENVI) for individual masked pixels MASK_pix.line - the line number (ENVI) for individual masked pixels MASK_pix.wvl - the wavelengths for individual masked pixels MASK_pix.Rrs - the Rrs spectrum for individual masked pixels (corrsponding to the wavelengths)
ENVI_GRID.wvl - wavelength corresponding to each masked 'layer' ENVI_GRID.layer - Rrs grid (layer) corresponding to the masked ESF