ETRI-Knowledge Sharing Plaform

ENGLISH

성과물

논문 검색
구분 SCI
연도 ~ 키워드

상세정보

학술대회 Automatic Number Plate Detection for Korean Vehicles
Cited 0 time in scopus Download 2 time Share share facebook twitter linkedin kakaostory
저자
윤호섭, 이홍창, 이재연
발행일
200911
출처
International Conference on Interaction Sciences: Information Technology, Culture and Human (ICIS) 2009, pp.1191-1195
DOI
https://dx.doi.org/10.1145/1655925.1656142
협약과제
09IC1800, u-로봇 HRI 솔루션 및 핵심 소자 기술 개발, 황대환
초록
We consider the task of detection of Korean vehicle number plates(also named license plates or registration plates in other countries). A system for number plate detection must cope with wide variations in the appearance of the plates. Few yeas ago, Korea government permits the new types of license plates with wide width as same as foreign country. It occurred that the new license plates detection algorithm was needed. We have developed a system that makes precise decision rules for detecting plate area using priori-knowledge about various plate types. According to the lighting condition, seed processing as adaptive Sobel edge is very important. In some cases, plate candidate areas are not selected. And then, we retry to Sobel edge detection with strong threshold value. Final detection results have shown that proposed detection algorithm is very useful. Copyright © 2009 ACM.
KSP 제안 키워드
Decision rules, Detection algorithm, Lighting conditions, Number plate detection, Sobel edge detection, Strong threshold, Threshold Value, license plate, seed processing