Abstract: This paper presents a volume rendering approach on current programmable consumer graphics hardware using ray-casting algorithm in the fragment shader of OpenGL2.0，which exploits the new features of modern consumer graphics hardware.In order to process the Large Volume Data Sets，in the first we directly load the volumetric data into texture memory，which has already defined in the form of 3D texture.In the second we utilize pre-integrated classification to improve the quality of rendering.At last，gradients are computed in real-time，which greatly improves the efficiency of ray-casting.The experiment results show that the approach can achieve high quality image and interactive frame rates.

Key words: ray casting, the programmable graphic hardware, fragment shader, 3D texture

摘要： 介绍了一种基于GPU（可编程图形处理单元）的快速实时光线投射算法。为满足大规模体数据集的绘制要求，利用当前GPU的新特性，直接将体数据作为纹理载入显存，采用预积分分类方法在GPU中对体数据进行重采样和分类，避免了计算机主内存与GPU纹理内存之间数据交换的瓶颈问题；利用硬件支持的三维纹理和片元着色器，实时计算每个体素的梯度，实现高质量的光照，保证高质量的图像绘制效果。实验结果表明该方法在医学三维数据场可视化中，能够实时、高效地生成高质量的交互式体可视化图像。

关键词: 光线投射, 可编程图形硬件, 片段着色器, 三维纹理