|
Article information
2015 , Volume 20, ¹ 1, p.63-74
Hmelnov A.E.
The MRG file format for compact representation and fast decompression of large digital elevation models
Purpose. A new file format MRG (MultiResolution Grid), which is designed for lossless storage and fast decompression of large integer raster images, is considered. The format developed allows to quickly fetch raster fragments of various levels of details on account of storing both the most detailed data and the data of lower resolution. Design ⁄ methodology ⁄ approach. The data of higher resolution are represented by storing the difference between actual data and their approximation from the corresponding lower resolution image block. For compression of differential integer sequences an original algorithm BCRL (Bit Count Run Length encoding) has been developed. The algorithm allows both to increase the data compression ratio and to reduce the data compression and especially decompression time (in comparison with the popular library ZLib). The BCRL decompression algorithm can be expressed by few lines of code: it just reads intervals of values of constant bit depths. The compression algorithm finds the close to optimal subdivision of the sequence of integer values into such intervals using the dynamic programming approach. Findings. The proposed data structure allows not only lower usage of additional memory for storing the additional low resolution raster images, but conversely, proposes how to decrease the overall memory use. The BCRL data compression algorithm can efficiently utilize the features of differential sequences. Originality ⁄ value. The file format developed is completely original and based on different principles. Compared to the known file formats, it can be used for lossless compression of integer raster images, such as MrSID, ECW, JPEG 2000. The proposed compression algorithm is significantly simpler, but still very efficient. A dynamic library has been implemented for reading data from the MRG files by external programs. The developed software can effectively work using a limited amount of RAM (down to 16 Mb).
[full text]
Keywords: data format, big raster data, lossless compression
Author(s):
Hmelnov Alexey Evgenievich
PhD. , Associate Professor
Position: Head of Laboratory
Office: Matrosov Institute for System Dynamics and Control Theory of Siberian Branch of Russian Academy of Sciences
Address: 664033, Russia, Irkutsk, 134 Lermontov str.
Phone Office: (3952) 45-30-71
E-mail: hmelnov@icc.ru
SPIN-code: 8041-3667 References:
[1] Dubinin, M. The SRTM data description and sources(2014), Available at: http://gis-lab.info/qa/srtm.html (accessed 23.09.2014).
[2] Jarvis, A., Rubiano, J., Nelson, A. et al. Practical use of SRTM data in the tropics —Comparisons with digital elevation models generated from cartographic data. Working document no. 198. Centro Internacional de Agricultura Tropical (CIAT), — Cali, CO. 32 pp. Available at: http://srtm.csi.cgiar.org/PDF/Jarvis4.pdf.
[3] Google Earth(2014), Available at: http://www.google.com/earth/ (accessed 23.09.2014).
[4] Vatolin, D., Ratushnyak, A., Smirnov, M., Yuckin, V. Metody szhatiya dannyh. ustroystvo arkhivatorov, szhatie izobrazheniy i video [Data compression methods. The structure of archivers, image and video compression]. Moscow: Dialog-MIFI, 2003. (In Russ.)
[5] Vatolin, D. Everything about the data compression(2014), Available at: http://www.compression.ru (accessed 23.09.2014).
[6] Salomon, D. A guide to data compression methods. New York: Springer-Verlag; 2002.
[7] Gailly, J.-L., Adler, M. ZLib home site(2014), Available at: http://www.zlib.net (accessed 23.09.2014)
Bibliography link:
Hmelnov A.E. The MRG file format for compact representation and fast decompression of large digital elevation models // Computational technologies. 2015. V. 20. ¹ 1. P. 63-74
|
