Line Integral Convolution of Vector Field for
I tried coding this up in Matlab to see if I could take some noise and generate an LIC over
2-D Unsteady Driven Cavity Problem for Reynolds 1000 and t ~ 25
a vector field. This corresponds to the t7 step for the Re 1000 example. The image is properly rotated.
Here is the noise that I gave to the Matlab code (It's just a random 280x280 matrix of color values from 0 to 256) with the output next to it. Here I used a filter length of 20. I also played with the kernel of convolution's constants some. There are three constants, c, d, and Beta where c and d represent the dialtion constants of the Hanning window and ripple functions respectively and Beta is the ripple function phase shift in radians. I tried d ~ 3*c to get the ripple function freq. to 2 - 4cycles per period for the window. I visibly couldn't tell if changing Beta affected much, and there are still some pretty bad arttifacts on the boundaries which I'm not sure if I'm treating properly.