====== Using SRTM Data in ROI_PAC====== The [[http://srtm.usgs.gov|SRTM]] dataset is one of the best elevation datasets to use in InSAR processing because of its near-global coverage (at 90m resolution) and 30m resolution for its U.S. coverage. - **Determine the spatial extent** of the SAR scenes that you will be processing. You can use EOLISA to determine the latitude and longitude of the corners of the scenes you have selected. Your DEM must completely cover the area of the SAR scene(s). - **Download the appropriate SRTM tiles.** SRTM data can be downloaded from the ftp site ftp://e0srp01u.ecs.nasa.gov. Here you can choose 1-arcsecond (SRTM1), 3-arcsecond (SRTM3), or 30-arcsecond (SRTM30) resolution. In general it is best to use the highest resolution version 2 ("finished") data available for your area. The data is available in 1-degree square tiles. **The SRTM tiles are named for the coordinates of the lower left (southwest) corner.** Download all of the tiles that overlap with your area. Even for a single SAR scene, 4 tiles are usually necessary. - **Unzip the files** using the UNIX ''unzip'' command, e.g.: \\ ''$ unzip N38W112.hgt.zip'' - **File architecture conversion** will be necessary if you are using an Intel processor, which assumes little-endian files. The SRTM data files from this FTP site are big-endian. Use the ''dd'' command to convert them: \\ ''$ dd if=//input file// of=//output file// conv=swab'' \\ Repeat for each ''.hgt'' file. - **Mosaic Tiles:** If your area spans more than one SRTM tile you will need to mosaic the tiles into a single file. This can be done using a Matlab script such as this: \\ \\ ''% read 30m SRTM DEMs \\ \\ fid=fopen('N37W122o.hgt'); \\ B1=fread(fid,[3601 3601],'int16'); \\ fclose(fid); \\ \\ fid=fopen('N37W123o.hgt'); \\ B2=fread(fid,[3601 3601],'int16'); \\ fclose(fid); \\ \\ fid=fopen('N38W122o.hgt'); \\ B3=fread(fid,[3601 3601],'int16'); \\ fclose(fid); \\ \\ fid=fopen('N38W123o.hgt'); \\ B4=fread(fid,[3601 3601],'int16'); \\ fclose(fid); \\ \\ % mosaic tiles \\ B=[B4' B3';B2' B1']; \\ \\ % delete overlapping row and column \\ B(3601,:)=[]; \\ B(:,3601)=[]; \\ \\ % replace bad data points \\ jj=find(B==-32768); \\ B(jj)=NaN; \\ \\ % view result \\ imagesc(B);'' \\ \\ Using the output figure you can check to make sure the files are in the right order. If so, use the following Matlab commands to save the file: \\ \\ ''% write output file \\ fid=fopen('bayarea.dem','w'); \\ fwrite(fid,B','int16'); \\ fclose(fid);'' \\ \\ You will of course have to replace the file names with those that you are using. This script can also be adapted to mosaic different numbers of files. - **Make a resource (RSC) file** for the DEM. ROI_PAC requires that all metadata for your DEM be contained in a resource file, which should have the same name as the dem with ''.rsc'' appended, e.g. ''bayarea.dem.rsc''. The format is as follows: <code> WIDTH 7201 (pixel width) FILE_LENGTH 7201 (pixel length) XMIN 0 XMAX 7200 (one less than WIDTH) YMIN 0 YMAX 7200 (one less than LENGTH) X_FIRST -123.000000000 (longitude of left edge) Y_FIRST 39.000000000 (latitude of top edge) X_STEP 0.000277778 (pixel size in degrees) Y_STEP -0.000277778 X_UNIT degrees Y_UNIT degrees Z_OFFSET 0 Z_SCALE 1 PROJECTION LATLON </code> Move the ''.dem'' and ''.dem.rsc'' files to the directory where you permanently store DEMs. The originally downloaded files and all other intermediate files may be deleted. You can now use your DEM to produce differential interferograms of your study area. * **Go to the next step:** [[makeraw|Prepare SAR data for processing]]