Implementing an algorithm for lossy compressing elevation data based on SPIHT

Web-based visual simulation systems have to handle very large, even huge volumes of elevation data sets. Data compression is an effective mean for reducing space storage and transmission time of these data sets on the Internet. A novel scheme for lossy compression of elevation data using Set Partitioning in Hierarchical Trees (SPIHT) is studied in this paper. An optimal linear predictor is used to identify and remove the redundant information among the neighboring samples in the first stage. Then by exploiting the prosperities of SPIHT, an effective method based on SPIHT is developed to encode the residual elevation data. The algorithm has been applied to data from USGS digital elevation model, and the compressing results strongly suggest that the proposed lossy data compression scheme provides a practical way for elevation data compression.