一种固体发动机光线投射体绘制加速方法

A Ray Casting Volumn Rendering Accelerated Method for Solid Motor

摘要:

固体发动机ICT断层图像构成的体数据场规模大、空体素比例小,现有的光线投射加速方法效果不明显,并且在其三维可视化故障诊断中,不是体数据场中的所有信息都对内部故障检测有意义,三维重建出所有体数据大大增加了计算量。为此,提出一种在固体发动机体数据分割基础上的光线投射加速方法。该方法利用分割结果,建立分割信息体数据场,结合现有的分层体数据结构构造方法,构建出基于分割信息的分层体数据结构。依据此结构,光线投射中自动识别用户定义的重建组成部分,将其他组成部分当作空体素跳过,动态调整采样步长,减少了采样点数目,提高了固体发动机体绘制速度。实验表明,该加速方法不仅效率高,且不影响显示效果。

Abstract:

The size of volume data field which be made up of solid motor ICT images is big, and empty voxels in the volume data field occupy little ratio. The effect of existing ray casting accelerated methods is not distinct. In 3D visualization fault diagnosis of large-scale war industry product, not all the information in the volume data field can help surveyor check out fault inside it. Computational complexity is greatly increased if all volume data is 3D reconstructed. So a new ray casting accelerated method based on solid motor segmented volume data was put forward. In the method, segmented information volume data field was built making use of segmented result. Consulting the conformation method of existing hierarchical volume data structures, hierarchical volume data structure on the base of segmented information was constructed. According as the structure, the construction parts which defined by user were identified automatically in ray casting. The other parts were regarded as empty voxels, hence the sampling step was adjusted dynamically, the sampling point amount was decreased, and the volume rendering speed was improved. Experimental results finally revealed the high efficiency and good display performance or the proposed method.