彩色三维体数据场的直接体绘制方法

Direct Volume Rendering from 3-D Color Data Sets

提出基于光线投射和三维纹理映射的彩色体数据成像算法,研究如何从每个体素的R,G,B三元组映射成不透明度值,即不透明度转换函数。首先把原始的RGB色彩空间转换成LUV色彩空间;然后以亮度分量的中心差分来近似估计法向量,并应用Phong光照模型进行着色,根据亮度分量及其梯度等信息计算不透明度值;最后合成、累积颜色。对美国数字人男子照相彩色体数据分别采用两种算法进行实验。结果表明:基于光线投射的彩色体数据算法成像质量较高,可以表现体表毛细血管等细微结构,但速度较慢;基于三维纹理映射的彩色体数据成像算法速度较快,但成像质量适中。

Two volume rendering algorithms based on the ray casting and the 3-D texturing are presented to visualize 3D color data sets. Both approaches are focused on the conversion of RGB tuples into opacity values called the opacity transfer function. Firstly, RGB color space is transformed into LUV color space, and the central difference of luminance is calculated to estimate the normal. Then, the volume is shaded by using the Phong illumination model. Finally the color and the opacity are composed. Experimental results on American visible human datasets show that the ray casting based algorithm can generate high quality images and display the capillary vessel and textures of the body surface, but the rendering speed is lower; the 3-D texturing based algorithm can achieve the interactive rendering speed but the image quality is moderate.