jie Li

A time-controlling terrain rendering algorithm

The paper proposes a time-controlling algorithm for large-scale terrain rendering, which can’t be efficiently dealt with by LOD technique. In this algorithm, the terrain is divided and organized by quad-tree structure. Each terrain patch is assigned a certain time according to the total rendering time given in advance. The multi-resolution levels rendered are determined by visual apperception. To solve the T-junction and popping, an approach of stitching the level boundaries and geomorphing are respectively performed on GPU. The algorithm guarantees that each frame is rendered in preset time independent of the terrain or the eye position. The slow or jerky phenomena during roaming, which are usually caused by unstable rendering frame rate, can be successfully avoided. This terrain rendering algorithm is demonstrated in a massive terrain flyover application. The experiment proves that this algorithm is feasible and efficient.

Method of motion data processing based on manifold learning

Due to the high-dimensionality of motion captured data which resulted in the complexity in motion analysis, a method of motion data processing based on manifold learning was proposed. Isomap, a classical manifold learning algorithm, was necessary to be improved and extended in this paper. A framework of motion data processing based on manifold learning was built to embed high-dimensionality data into low-dimensionality space. It simplified the motion analysis, and in the same time preserved the original motion features. In order to solve the inefficiency of processing large-scale motion data, Sample Isomap (S-Isomap) algorithm was proposed. Experiments proved that approximate embeddings of motion data computed by S-Isomap were average 10 times faster than by Isomap, while 10% frame samples were selected.

Smooth Schedule of Large-scale Terrain Visualization from External Memory

The paper proposed a smooth schedule algorithm to render terrain data out-of-core. The terrain data was organized as a quad-tree with Z-order filling curve, which preserved the data continuity and improved the schedule efficiency. A schedule strategy basing memory allocation was designed to support the stable frame rate terrain rendering. Controllable schedule area, instead of the view frustum culling, was applied to achieve the balance between the in-core memory requirements and scheduling time. A pre-estimated approach was adopted to smooth the data exchange. The algorithm achieved a smooth schedule and successfully avoided the slow or jerky phenomena, usually caused by inefficient and unsmooth schedule. By comparing with traditional methods, the algorithm proposed is feasible and efficient in out-of-core terrain visualization.