Fast Visualization of Stress and Strain Tensor Fields on Streamsurfaces with LIC Texture

摘要

Stress and strain tensor fields are specific second-order symmetric tensor fields, and the visualization remains a challenging task which is expected to help scientists and engineers get insights into the fields themselves and the related physical problems. In this paper, we propose to visualize fields on streamsurfaces with LIC texture. Streamsurfaces are adopted to display one of the three eigenvector fields of a tensor field which is orthogonal to them and were presented originally by some researchers in the case of diffusion tensor fields. To represent the other two tangent eigenvector fields upon streamsurfaces also, we generate dense LIC texture images for them by the image space based visualization method. In consideration that for stress and strain tensor fields, the signs of the eigenvalues have physical meaning: "+" indicates expansion and "-" means compression, we encode this information simultaneously within the same LIC texture images by using sparse noise images, as opposed to white noise images, as input and employing some free parameters of sparse noise images. Thus we achieve a complete visualization of all of the three eigenvector fields of stress and strain tensor fields upon streamsurfaces including the signs of the tangent eigenvector fields. Additionally, to acquire high computational speed, we generate the LIC texture images for the two tangent eigenvector fields with CUDA of GPU.