返回Google圖書搜尋

Human-like algorithm for 2D Delaunay triangulation

出版	Faculty of Electrical Engineering and Computer Science, Laboratory for geometric modelling and multimedia algorithms, 2003
URL	http://books.google.com.hk/books?id=YUbNOAAACAAJ&hl=&source=gbs_api

註釋The report introduces a new algorithm for constructing a 2D Delaunay triangulation. It bases on the sweep-line paradigm, which is combined by a local optimisation criterion - a caracteristic of the incremental insertion algorithms. The sweep-line status is represented by so-called advancing front, which is implemented as a hash-table. To prevent the construction of tiny triangles, which would be generated by a direct implementation of the sweep-line approach, heuristics have been introduced. The algorithm has been compared with other popular Delaunay algorithms: improved Gubais, Knuth, and Sharir's incremental algorithm, Žalik and Kolingerova randomised incremental algorithm using two-level uniform space subdivision, Muecke's et al randomised incremental walking algorithm, Lee and Schachter's divide and conquer algorithm, and Fortune's sweep-line algorithm. Schewchuk's implementation of the last three algorithms has been used. The proposed algorithm is for at last 400% faster that fastest algorithm among tested ones. Beside that, the algorithm does not use a lot of memory for supporting data structure, it is easy to understand and simple to implement. All of that align the proposed algorithm among the best algorithms for constructin Delaunay triangulation.