Kazunori Yamamori1*, Hiroshi Honda2 and Jun-ichi Hasegawa3
1Information Processing Center, Mie University, 1515 Kamihama-cho, Tsu-shi, Mie 514-8507, Japan
2Densi System Co., 1-22 Koubai-cho, Shouwa-ku, Nagoya-shi, Aichi 466-0031, Japan
3Graduate School of Computer and Cognitive Sciences, Chukyo University, 101 Tokodachi, Kaizu-cho, Toyota-shi, Aichi 470-0393, Japan
*E-mail address: email@example.com
(Received October 22, 2001; Accepted May 2, 2002)
Keywords: Map Information Processing, Road Deformation, Street, Parallel Transformation
Abstract. This paper presents a new method deformation of road networks developed for the automated generation of rough guide maps seen in advertisements and leaflets. To address the same problem, we have already developed a parallel method which arranges road segments (edges) by quantizing the direction of each edge in parallel and iteratively. However, this previous parallel method has a problem in that it cannot always obtain good results from a global viewpoint because road transformation is performed edge-wise. In the new parallel method, road transformation is performed street-wise. As a result, the new method can greatly transform a cluster of edges from a global viewpoint and streets can be transformed differently according to their shape and direction. In the experiment using 88 real road maps, all results obtained using the new method were more comfortable than those obtained by using the previous one.