(A) real-time visualization algorithm of large terrain data using wavelet-based compression

Abstract

Visualization of very large terrain data in network environments is an important issue in GIS, computer games, and military applications. For real-time visualization, however, the problems due to the narrow bandwidth of the network and the limited computational power are to be solved. Although the previous algorithms based on a qaudtree structure can visualize terrain data in real-time, they require too large amount of data to be transmitted over the network. Therefore, we propose a real-time terrain visualization algorithm combined with wavelet-based compression. Our approach introduces a wavelet-based compression scheme having a multi-resolution structure to compress a large terrain data, thereby, handle it interactively. Also, for visualization, a surface mesh model is updated in real-time by using wavelet coefficients. To achieve this real-time mesh updating, a new mesh approximation method using restricted quadtree triangulation is designed on the basis of wavelet coefficients representing surface complexity. By combining the wavelet compression and wavelet-coefficient-based visualization techniques, the compressed data can be decoded and visualized interactively. Simulation results show the proposed algorithm can change the viewpoint and display very large terrain data in real-time with only several tens of kbits per second. Hence, the proposed algorithm is appropriate for the applications in network environments that usually have a limited narrow bandwidth and a low computational power.