The Research And Implementation On High Dynamic Range Volume Rendering

On the conventional display device, direct volume rendering of high dynamic range data is usually limited by the dynamic range of display devic. Relatively narrow dynamic range of display device and usual 8-bit of color resolution on the monitor have become the bottleneck of visualization for high dynamic volume data which covers with a large range and distributes very unevenly. In this paper, we focus on the approach for high dynamic volumetric visualization based on the tone mapping operator and the method utilizing histogram equalization for transfer function design. The main contribution of this paper can be outlined as the following three aspects.Firstly, for high dynamic range three-dimensional data set, the distribution of density is very uneven. In the process of volume rendering, the usual colormap can map the data value onto color space, but it always makes the result picture difficult to distinguish the difference of intensive part of the data. We propose a colormap reproduce algorithm for high dynamic range volume data, which can adjust pre-set color map according to the characteristics of volume data. Experimental results show that the reproduced colormap can better distinguish the difference of the intensive part of volume data.Secondly, for large scale high dynamic range volume datasets, it is usually not so easy to achieve the desired rendering results by traditional transfer function methods. We propose a novel transfer function design method for high dynamic volume data. The method adopts a tone mapping operator to the transfer function. The tone mapping operator maps high dynamic range data to colormap based on dataset's global histogram. Experimental results show that rendering images can present much more details when the method is used for our seismic datasets.Finally, for high dynamic range float volume data with limited information about their content, deriving transfer functions using mathematical properties is a difficult task. We propose an opacity histogram equalization based transfer function design method for such volume data. Experimental results show that this approach can better highlight the rich detail levels of high dynamic range volume data, and also greatly improve the efficiency of transfer function design.All of these methods have been applied in our seismic dataset visualization system, and achieved good results.