http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
히트맵을 활용한 딥러닝 기반 주차선 검출 모델에 관한 연구
백종원(Jongwon Baek),정현석(Hyeonseok Jung) 한국자동차공학회 2023 한국자동차공학회 부문종합 학술대회 Vol.2023 No.5
Recently, as autonomous driving cars have been developed and commercialized, numerous studies are being conducted to secure higher levels of autonomous driving technology. Autonomous driving cars require not only autonomous driving but also autonomous parking technology. The current autonomous parking assist technology assists the driver by recognizing the parking space and calculating the parking trajectory by combining cameras and ultrasonic sensors attached to the vehicle. While ultrasonic sensors can detect empty parking spaces between walls, pillars, and vehicles, cameras have a much greater reliance in detecting empty parking spaces with nothing around them. In addition, the performance of parking line recognition is essential when attempting to park properly aligned in the parking space. Currently, there is no international standard for the types or specifications of parking lines, and there exist various indoor and outdoor parking environment types with different types of parking lines in our society. Therefore, an excellent performance parking line detection model is needed. For these reasons, this paper proposes a deep learning-based parking line detection model utilizing a heatmap.