Parallel Rendering Solid Polylines And Single-colored Polygons On The Terrain Surface

In the early stage of researches on three-dimensional geographic information system (GIS), photorealistic rendering was gained more attention. Undulating terrain can be expressed by using digital elevation models (DEMs), and details of terrain surface can be expressed by using digital orthophoto maps (DOMs). However, there are too much details in the DOMs, which will make observers be easy to get lost in the3D scene because of the lack of abstract information.Two-dimensional vector features are the abstraction of the objective world, which can overview and express spatial phenomenon or spatial characteristics. Rendering two-dimensional vector features on the photorealistic performance terrain surface will strengthen the ability of the representation of spatial phenomenon in3D scene, which unions realistic and abstract. Theories and methods researches on this area will not only enhance expression of terrain surface, but also will affect the interactive operation, and even enhance spatial analyst ability and performance modes of3D GIS.Based on the former fact, a method of parallel rendering solid polylines and single-colored polygons on the terrain surface is proposed under support of National Natural Science Foundation of China (41371365). This method involves rendering algorithms separately for solid polylines and single-colored polygons based on the screen-space computation, and is achieved by heterogeneous computing. The main contents and conclusions are as follows:(1) Creating appropriate spatial indexes for vector data. Creating quad-tree spatial indexes for vector data in the process of vector data processing and organization. In order to improve the searching efficiency, cutting polylines into fragmentations before indexing is necessary. Spatial indexing effectively reduces the time consumption on searching features and improves the operation efficiency of the method.(2) Designing rendering algorithms for solid polylines based on screen-space computation. Calculating each pixel’s extent in the coordinate of vector data accurately, then designing fast algorithm on judging the relationship of pixel’s extent and polylines. Furthermore, detecting the silhouettes of rendering images to remove the wrong shading pixels, which makes rendering results correctly. Finally, widening the polylines with morphology and achieving the polyline’s gradual color changing.(3) Designing rendering algorithms for single-colored polygons based on screen-space computation. Single-colored polygons can be classified into two categories, one category is the polygons without outlines, and the other is the polygons with outlines. For polygons without outlines, shading the pixels by judging whether their centers fall in polygons or not. While for polygons with outlines, shading inner surface firstly and then rendering their outlines based on the polylines rendering algorithm, which makes outlines connect with inner surface.(4) Parallelized the rendering method based on heterogeneous computing. Dividing the method procedure into serial processing section and parallel processing section. Providing the serial processing section to CPU, which is good at dealing with complex logical operations. Meanwhile, providing the parallel processing section to GPU, which is designed to deal with parallel computing. Heterogeneous computing improves rendering efficiency.Through testing the proposed method, rendering image result is correct and the operating efficiency is satisfied.