Abstract: An algorithm is proposed to develop a way of high-quality real-time three dimension surface reconstruction for digitized human color images.Firstly，segment the 3D surface point sets of single organ using an interactive segmentation platform.Then，according with the gray gradient of volume data which has been filtered，calculate the normal vector of every surface point.Last，using surface points with color to describe the organ surface，and display the surface with OpenGL interface of display card.On hardware platform of personal computer，execute the reconstruction of liver and lung for the color images of visible human of USA，and the rendering speed excesses 25F/s.The algorithm can realize the real-time rendering for high resolution color volume data.

Key words: surface rendering, color images, digital human

摘要： 对数字人彩色照片数据进行高质量的实时三维表面重建提出了一种新算法。该算法利用交互分割平台提取出彩色体数据中单个器官的三维表面点集，再根据对二值体数据滤波后的灰度值计算灰度梯度估算得到表面点的法向量。最后用带颜色的表面点来描述器官的三维表面，利用显卡OpenGL接口对表面点集进行三维显示。在微机环境下对美国数字人照片数据集中的肝脏和肺部两个器官进行了三维重建，在保证图像质量的前提下重建速度超过25帧/s。提出的算法能对高分辨率的彩色体数据进行高质量的实时三维表面重建。

关键词: 表面重建, 彩色照片, 数字人