Title :
A new edge extracting method for vehicle navigation
Author :
Haifeng, Yang ; Mäkelä, Hannu
Author_Institution :
Lab. of Electr. & Autom. Eng., Tech. Res. Center of Finland, Espoo, Finland
Abstract :
A new edge detection technique is described, which can be used in vehicle navigation and other fields. These edges are one-pixel wide and without break points. The vanish point as well as the guide line are extracted in the navigation system based on knowledge and inference. The edge detection technique described is based on new definitions and a new chain code definition is developed. The advantages of these definitions are: fast searching, connecting break points and reducing memory space. An algorithm is developed based on these ideas and definitions as well as analyses of cost function, reliability and simplicity. This algorithm is convenient for fast searching and following the edges. It can be used in real-time control.<>
Keywords :
computer vision; edge detection; encoding; image recognition; navigation; vehicles; break points connection; chain code definition; edge detection; image recognition; inference; vanish point; vehicle navigation; Automation; Automotive engineering; Computer vision; Image edge detection; Laboratories; Navigation; Pattern recognition; Pixel; Roads; Vehicles;
Conference_Titel :
Advanced Robotics, 1991. 'Robots in Unstructured Environments', 91 ICAR., Fifth International Conference on
Conference_Location :
Pisa, Italy
Print_ISBN :
0-7803-0078-5
DOI :
10.1109/ICAR.1991.240419
