Dear OpenJUMP developers, Last month, I posted a note to this mailing list that introduced the Tinfour open-source library for processing Delaunay Triangulations. I received a favorable response and a few interesting ideas that I am thinking about incorporating into my code. I wanted to start by thanking everyone for the kind words.
I have one additional project that might be useful to the OpenJUMP community. Over the last couple of years, I have investigated techniques for lossless data compression for gridded geospatial data sets. I’ve implemented some of them in Java and posted them at the Gridfour software project at https://github.com/gwlucastrig/gridfour . Because the techniques compress data without a loss of precision, they are suitable for applications that require persistent storage of data or transmission across bandwidth-limited channels. And while none of the techniques I use are truly novel, they do improve on the results currently available from GeoTIFF and NetCDF files. If you are interested, feel free to use the code as is or to repurpose it for your own applications. If you would like more information, the following links are a good place to start: 1. General Compression: https://gwlucastrig.github.io/GridfourDocs/notes/GridfourDataCompressionAlgorithms.html 2. Lossless Floating-Point Compression: https://gwlucastrig.github.io/GridfourDocs/notes/LosslessCompressionForFloatingPointData.html 3. Enhanced compression using optimal predictors: https://gwlucastrig.github.io/GridfourDocs/notes/CompressionUsingOptimalPredictors.html 4. Multi-threaded techniques to speed storage and retrieval: https://github.com/gwlucastrig/gridfour/wiki/GVRS-Using-Multiple-Threads-to-Speed-Processing So far, my results are modest. I am working on improvements. Here are some values in bits-per-data-value for well-known Digital Elevation Models (DEMs). The SRTM elevations are given in individual files covering a 1-degree area which I’ve identified by a common name for reference. In general, integer data compresses much more readily than floating point. And grids with denser cell spacing compress better than coarser grids. 1. ETOPO1 (integer, global, 1 minute grid spacing): 3.65 bits/value 2. GEBCO 2021 (integer, global, 15 second grid spacing): 2.86 bits/value 3. SRTM Padua, Italy (integer, 1 second grid spacing): 2.23 bits/value 4. SRTM Sioux Falls, SD, U.S. ( integer, 1 second grid spacing): 1.74 bits/value 5. USGS 3DEP State College, PA, U.S. (single precision floating-point, 1/3 second grid spacing): 17.2 bits/value Finally, I am always looking for new ways to improve my techniques. If anyone has suggestions, I would be happy to hear them. Thanks, Gary
_______________________________________________ Jump-pilot-devel mailing list Jump-pilot-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/jump-pilot-devel
