Mode 1:
ACORN has a relatively user-friendly interface but the difficulty is finding the necessary input files from the information provided by the imagery people. In this case that was JPL at NASA. The AVIRIS imagery dataset comes with a set files that do contain all the information you need for ACORN. It all had to be loaded into excel, changed slightly and then saved in a tab-delimeted ascii format, or simply read off of the file and typed into the ACORN control screen.
The Header File (.hdr) next to each image scene contains the byte-interleave (bip, IEEE for AVIRIS 2002 datasets.) as well as the number of bands, lines, and samples. That's 224 bands for AVIRIS, but the lines and samples may change with each image.
The spectral calibration file (.spc) comes up as some sort of PCKS Certificate according to the people at MSN who like to think they either know or can anticipate everything. Open it in excel as ascii and keep only the first two columns - the band number and Full-Width-Half-Max (FWHM) of the appropriate Gaussian function. Save As tab-delimited ascii and its ready for ACORN.
The GAIN file (.gain) can be opened in word, but use excel again because you have to reformat the file. first you only need the first column so delete the second one (band number). Now getting the units to match was tricky, but as it just turns out the conversion factor is ten for the different units. [10 watts/meter^2/micron/steridian = 1 microwatt/centimeter^2/nanometer/steridian] ACORN wants the one that starts with watts and AVIRIS comes in the one that starts in microwatts. I also had to take the inverse of the AVIRIS values because they were all geared up for dividing while ACORN wants to multiply. So remember first column only inverted and multiplied by ten!
There is no given offset. I believe that the AVIRIS offset for 2002 was 0. That means I made a single column ascii text of 0 with 224 rows. 224 X 0.
The average Lat/Long, and GMT time are easily found at aviris.jpl.nasa.gov under status 2002. I averaged them and used that info for the whole set of imagery scenes. I believe that this information is also somewhere in their giant navigation file (.NAV), which has some headers, lots of confusion, a few mis-alignments, and about 7,000 columns of information. I didn't count the rows personally. Easily retrieved out of the NAV file is the altitude information, which is clearly labeled and all in good alignment. I took the average of the whole column. Knowing the elevation of the image is your own responsibility. Use the Web if you don't already have something better.
I did some testing and in the end decided to supress all three artifacts. Pickier people who are good at IDL may decide to do otherwise. Or I guess if you were in a hurry, each artifact selected makes it take longer, but I was still looking at 30min for a 1GB image on a 2400MHZ pentium something or other and I was bringing it off of the server and putting it back during processing. not bad, eh?
You have to put in something for the Fixed Water Vapor and Atmospheric Visibility but I didn't collect that information myself and ACORN with improve your guestimate rather well if you the check the boxes for each.
There are a number of files outputed along with the calibrated image. Diag1 and Diag2 are full of good information. The ECO file and IN file record what your input parameters were.
Have a nice day.
Last updated May 14, 2003