The Reaserch And Implementation Of Triangle Rasterization And The Anti-Aliasing Algorithm On Graphics Processing Unit

The performance of rasterization and anti-aliasing have a great influence in the GPU.In order to meet people's requirements on the performance of the chip,it is imperative to improve the rasterization and anti-aliasing technology.Rasterization as the most basic module in image rendering,mainly carries the task of converting primitives from geometric information into pixel segments.Anti-aliasing aims to solve the problem of signal distortion in the rasterization process,so that the step-like sawtooth appearing on the edge of the image will be improved and the image will have more visual effects.In image drawing,the main object is the drawing of polygons.And in practice,it is divided into multiple triangles to draw.Therefore,the research on rasterization and anti-aliasing of triangles is particularly important.The following is the research in this paper:(1)The triangular rasterization algorithm and propose optimization has been studied.An improved edge function scanning algorithm based on Bresenham algorithm has been proposed.What's more,the pixel scanning mode has been changed from a single scanning line scanning mode to parallel scanning,and four scanning lines have been taken as an example for illustration.After increasing the scanning line for scanning,drawing efficiency has been greatly improved.The property interpolation algorithm and perspective correction algorithm for this rasterization algorithm have been studied.(2)Triangular anti-aliasing algorithm has been studied and improved.Aiming at the analysis and comparison of the anti-aliasing methods based on hardware implementation of a variety of triangles,the scope of application has been discussed,and an optimization and improvement have been proposed for the multi-sample anti-aliasing algorithm of the edge equations.The core of the algorithm is to divide a pixel into a plurality of sub-pixels,and determine the number of sub-pixels located in the triangle by determining the positional relationship between sub-pixels and triangles,and calculate the color value of the pixel.The improving judgment method,which is using the iterative parameters instead of the original calculation unit,has reduced unnecessary operations and improved the efficiency of the algorithm.(3)The realization of the hardware circuit has been completed according to the improved algorithm principle.The rasterization process has been uniformly described,and rasterization and anti-aliasing have been designed in parallel to reduce the amount of calculations.A six-stage pipeline design method has been used in the rasterization module,they are vertex attribute reception level,coordinate parameter establishment level,attribute parameter establishment level,base point attribute correction level,first pixel attribute interpolation level,and horizontal attribute interpolation level.The specific implementation flow and the core judgment circuit unit have been given.(4)The improved rasterization and anti-aliasing algorithm have been verified.A simulation platform has been set up,and verification items have been planned according to the design principle.Module-level simulation was completed and the correctness of the improved algorithm was verified.The main verification directions are the correctness of triangle drawing,the drawing of special triangles,the quality of image anti-aliasing,collection of pixel pass rate,vertex pass rate,and algorithm rendering efficiency.After using the parallel scanning method to rasterize the image,the efficiency of the drawing of the same triangle has reduced from 52.9 ns to 19.8 ns.