The url is https://hdl.handle.net/1912/67567 and the DOI is 10.26025/1912/67567

SAR image data are presented in raw and processed format.

The raw SAR modulation data have the extension *.dat and are in unformatted binary 4-byte floating point big-endian.  Imagery has been mapped onto a regular latitude/longitude grid with pixel spacing of about 15 m. 

In MATLAB it would be read with the following code 

    fid = fopen(dfile, 'r', 'b');
    tmpdata = fread(fid,'float32');
    image = reshape(tmpdata,[nl nc]);
    fclose(fid);
    clear tmpdata

where dfile is the filename of the data file, nl is the number of lines in the data file, and nc is the number of columns. For these files, nl=nc=2000. The first index is longitude and the second index is latitude.

The values in the SAR modulation data files are in units of amplitude so the data needs to be squared to get intensity (or energy).  They are stored in amplitude because that has less of a dynamic range and makes prettier pictures.  However SAR modulations are usually defined as (I-mean)/mean where I is image intensity (so the square of what is in the data file) and mean is the mean intensity value over the ocean.  Thus the data needs to be squared before calculating any modulation values or performing any radar image processing.      

For each raw data file, an associated ascii header file that contains the information about each image.  These all have extension *.exhdr give the information about the data such as number of lines, number of columns as well as the starting,ending latitude and longitude values for the grid.


Processed data files of SAR image intensity are provided in MATLAB file foramt (*.mat). The image intensities are normalized by calculating the mean ocean image value over a common region of mostly open ocean. The mean from this reference region was subtracted from the data, and the data were linearly mapped onto
[0,255] by setting -0.30 to 0, +0.30 to 255, and clipping at the ends.

Variables in the processed .mat files:
dn = time of image, Matlab datnum (UTC)
lat = latitude coordinates of image
lon = longituded coordinates of image
C = normalized image intensity, [0 255]