The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Download
Publications Copernicus
Download
Citation
Articles | Volume XXXVIII-4/W25
https://doi.org/10.5194/isprsarchives-XXXVIII-4-W25-7-2011
https://doi.org/10.5194/isprsarchives-XXXVIII-4-W25-7-2011
30 Aug 2012
 | 30 Aug 2012

A DTM MULTI-RESOLUTION COMPRESSED MODEL FOR EFFICIENT DATA STORAGE AND NETWORK TRANSFER

L. Biagi, M. Brovelli, and G. Zamboni

Keywords: Internet/Web, LIDAR, Modelling, Archiving, Compression, DEM/DTM, Algorithms, GIS

Abstract. In recent years the technological evolution of terrestrial, aerial and satellite surveying, has considerably increased the measurement accuracy and, consequently, the quality of the derived information. At the same time, the smaller and smaller limitations on data storage devices, in terms of capacity and cost, has allowed the storage and the elaboration of a bigger number of instrumental observations.

A significant example is the terrain height surveyed by LIDAR (LIght Detection And Ranging) technology where several height measurements for each square meter of land can be obtained. The availability of such a large quantity of observations is an essential requisite for an in-depth knowledge of the phenomena under study. But, at the same time, the most common Geographical Information Systems (GISs) show latency in visualizing and analyzing these kind of data.

This problem becomes more evident in case of Internet GIS. These systems are based on the very frequent flow of geographical information over the internet and, for this reason, the band-width of the network and the size of the data to be transmitted are two fundamental factors to be considered in order to guarantee the actual usability of these technologies. In this paper we focus our attention on digital terrain models (DTM's) and we briefly analyse the problems about the definition of the minimal necessary information to store and transmit DTM's over network, with a fixed tolerance, starting from a huge number of observations. Then we propose an innovative compression approach for sparse observations by means of multi-resolution spline functions approximation. The method is able to provide metrical accuracy at least comparable to that provided by the most common deterministic interpolation algorithms (inverse distance weighting, local polynomial, radial basis functions). At the same time it dramatically reduces the number of information required for storing or for transmitting and rebuilding a digital terrain model dataset. A brief description of the method is presented and comparisons about the accuracy and data-store compression obtained with respect to other interpolators are shown.