基于特征的矢量场自适应纹理绘制

针对3D流场纹理可视化方法面临的遮挡问题,提出了一种基于流场特征模糊提取的自适应稀疏纹理绘制方法.该方法建立在一种可扩展的模糊流场特征区域描述与提取算法基础上,首先通过采用模糊理论对流场特征区域进行描述,建立了流场特征的测度规则并得到相应特征向量,然后基于特征向量偏差给出了最小平方和误差准则下的模糊测度隶属度的确定方法.根据得到的流场特征模糊隶属度场,生成可突出流场特征信息的自适应高斯噪声以进行纹理LIC卷积.同时针对纹理方法难以明确表示3D流场具体方向的缺陷,提出了两种冷暖光照方法,通过在LIC卷积过程中采用冷暖光照方法突出流场具体方向显示,有效解决了该问题.实验表明,本文方法相对于传统特征提取方法更具有可扩展性.同时自适应技术的采用,有效缓解了3D流场可视化存在的遮挡和混淆现象.而冷暖光照处理的加入,则解决了纹理方法对流场具体方向表现上的缺陷.

Feature-based adaptive texture visualization for vector field

An adaptive sparse texture rendering method is proposed to solve for occlusion effects when visualizing 3D flows, building on an extensible fuzzy feature extraction approach. First, the flow feature is described by fuzzy theory and rules for some typicM features are obtained. The significance value for each voxel is then calculated by a clustering method under the minimum square-sum rule. An adaptive Gaussian noise field is obtained from the significance field by a noise generation process, and is used as the input for the LIC convolution process. We also present two cool/warm-illumination-based approaches to overcome the shortcomings of texture-based visualization methods, which are usually unable to represent the flow direction. The experiments show that our method can effectively extract the typical flow feature region and can be extended to other flow features easily, and the adaptive technique used lessens the occlusion effects significantly. Furthermore, the main disadvantage of the texture-based method, that is, the direction representation problem, can also be solved by the proposed cool/warm illumination methods.