Presentation slides can be found here: https://app.box.com/v/textureshading/file/1648840509141
Texture Shading: A Cartographic Algorithm for Showing Terrain Relief
Visualization of mountainous topography is commonly done using a shaded relief algorithm that combines an imaginary light source with a Lambertian reflectance model. The resulting images are intuitive to understand, but they also suffer certain drawbacks. Among these is anisotropy, or directional dependence, whereby the choice of lighting direction favors certain terrain features at the expense of others. Another common issue is lack of visual hierarchy, since small terrain features tend to clutter the image and obscure larger landforms that are actually more prominent. Most attempts to mitigate these problems have stayed within the general paradigm of illumination models.
A relatively recent algorithm called texture shading attempts to address the shortcomings of conventional shaded relief by taking a different approach, not derived from a lighting model. The underlying mathematics uses a fractional Laplacian operator adapted to work on discrete gridded data (in this case, a digital elevation model). The algorithm is isotropic, and scale-invariant in a particular sense, which gives a clear visual hierarchy to the images, and it highlights the network of ridges and canyons in the terrain. Fractional derivatives appear to be well-suited to terrain data due to the fractal nature of many topographic structures over a wide dynamic range.
The presentation will cover the motivation and goals of the algorithm, the implementation and mathematical details, example results, potential applications, and future work.