Fast Image‐Space Silhouette Extraction for Non‐Photorealistic Landscape Rendering

Abstract

Landscape illustration, a core visualization technique for field geologists and geomorphologists, employs the parsimonious use of linework to represent surface structure in a straightforward and intuitive manner. Under the rubric of non-photorealistic rendering (NPR), automated procedures in this vein render silhouettes and creases to represent, respectively, view-dependent and view-independent landscape features. This article presents two algorithms and implementations for rendering silhouettes from adaptive tessellations of point-normal (PN) triangles at speeds approaching those suitable for animation. PN triangles use cubic polynomial models to provide a surface that appears smooth at any required resolution. The first algorithm, drawing on standard silhouette detection techniques in surface meshes, builds object space facet adjacencies and image space pixel adjacencies in the graphics pipeline following adaptive tessellation. The second makes exclusive use of image space analysis without referencing the underlying scene world geometry. Other than initial pre-processing operations, recent advances in the OpenGL API allow implementations for both algorithms to be hosted entirely on the graphics processing unit (GPU), eliminating slowdowns through data transfer across the system memory bus. We show that both algorithms provide viable paths to real-time animation of pen and ink style landscape illustrations but that the second demonstrates superior performance over the first.

Concurrent Drainage Network Rendering for Automated Pen‐and‐ink Style Landscape Illustration

“…Sophisticated graphics programs can function with only vertex and fragment shading stages but OpenGL versions 4.0 and later give the programmer the means for finer control of the graphics pipeline (Wikipedia ). Mower () and this article both use tessellation control and evaluation shader stages to create Bezier surfaces for evaluation at finely tessellated mesh vertices. Doing so produces analytical surfaces with less surface variability for subsequent feature extraction than do non‐smoothed surfaces.…”

“…Since patches are processed independently of one another, an algorithm must employ a mechanism to ensure visual continuity between rendered patches. Mower () and the algorithm described in Section 2.3 use point–normal (PN) triangle patches to create a smooth analytical surface from a triangulated regular grid elevation mesh (Vlachos et al. ).…”

“…Since we are not concerned directly with silhouettes in this article, the reader may refer to Mower () for a complete description of that process.…”

“…This article introduces an algorithm for the automated real‐time rendering of creases – linework that helps an observer understand local surface curvature independently of his or her point of view – at speeds exceeding those required for animation. It also presents an implementation of the algorithm in the context of a comprehensive pen‐and‐ink style rendering system that includes a technique for fast silhouette rendering introduced in Mower (). Silhouettes are linework elements that mark the edge of a feature superimposed on a background feature or the sky from the observer's point of view (Figure ).…”

“…Current graphics platforms largely permit the implementation of earlier sequential NPR algorithms in ‘legacy’ or compatibility profile modes, but doing so generally ignores the advantages of their massively‐parallel architectures. Mower () introduces algorithms and an implementation for rendering silhouettes at animation speeds. Designed for modern GPUs, its implementation achieves animation frame rates on contemporary GPUs by limiting data transfer between main system memory and the GPU, by exploiting built‐in tessellation stages of the programmable pipeline, and by keeping synchronous inter‐core GPU communications to a minimum.…”