The Algorithm Research On LIC Based Visualization In Vector Field And It's Application In NPR

The visualization of vector fields has been a challenging issue in the Visualization Scientific Computing. It can directly display the motion of the vector field by graphics and images, and make people explain and understand the objective law of abstract science data visually. Line integral convolution (LIC) is the most classical and texture-based method for 2D vector field visualization. Vector field visualization based on LIC method, and many other improved LIC methods have been researched in this thesis firstly. Then we propose an iterative and adaptive integral length fast LIC (IAFLIC). This method could be applicable for any two-dimensional images, and it can be combined with the non-photorealistic rendering technique for stylized rendering results. The main contents of this thesis are as follows:Firstly, the thesis introduces the LIC method and the Fast LIC method, the uniform integral length is used in these two methods for vector field visualization. If an inappropriate integral length is used for a complex image, the visualization effects will be turbid, incoherent, and detail information lost. In view of this situation, we adopt a statistical technique to calculate integral length for every point in the vector field; it receive a short integral length for those complex regions for preserving the detail information of image, and a long integral length is obtained for those homogeneous flat regions to maintain the continuity of the visual effects. We introduce the accelerated theory of Fast LIC method for improving the efficiency of the algorithm in the process of the implementation, referred as: AFLIC method.The study showed that the effects from both of the original LIC method and our AFLIC method have some disadvantages, such as: blurred image, color unstable and so on. Then we adopt iterative AFLIC method, denoted as IAFLIC, to amend these disadvantages, that is, the output image from the previous convolution was used as the input texture of the next convolution. The experimental result shows that we can improve the image definition and color stability by this method, and the results are satisfied.Finally, we introduce the IAFLIC method to any 2D images for their texture visualization. We can obtain a stylized method for edge map extraction by filtering the edge vector field with an anisotropic Difference of Gaussian filter. It can extract a set of coherent, smooth, and stylized lines from 2D images, and these lines effectively capture and convey important shapes and details in the image. In contrast to conventional edge detector canny method, the edge map generated by proposed algorithm has more hand-painting characteristic. Moreover, many other abstract results, such as painting pictures and pencil drawing, could be obtained by combining the IAFLIC method with Non-Photorealistic Rendering technique.