카메라와 라이다의 센서융합에 의한 차선 및 과속방지턱 검출 방법

다국어 초록 (Multilingual Abstract)

In this paper, we propose a method to detect lane and speed bump that can be detected on the road using sensor fusion. To prevent this, it is necessary to detect the speed bump and determine its position. Further, the detection of the lane is a necessary factor for the accurate running of the vehicle. To detect this, we used a fusion of a lidar sensor to obtain accurate environmental information and a camera sensor to obtain various environmental information. We propose a multi - classifier method using fusion data using the pattern of speed bump and sensor fusion. First, the primary classification is performed through the image pattern of the speed bump, and then the distance information and reflectivity information of LiDAR data are projected on the image pattern to extract fusion data and proceed to secondary classification. Finally, by using the color information of the speed bump, the third speed classification is performed to detect the speed bump. Experimental results show that the triple classifier improves the accuracy of speed bump detection and distinguishes it from crosswalks with similar patterns.

목차 (Table of Contents)

• I. 서 론 1
• 1.1. 연구 배경 1
• 1.2. 기존 연구 4
• 1.3. 연구 목적 9
• 1.4. 논문 구성 12
• II. 시스템 구성 14
• III. 전처리 과정 16
• 3.1. 카메라 특징 추출 16
• 3.2. 라이다 특징 추출 19
• 3.3. 카메라-라이다 캘리브레이션 22
• IV. 과속방지턱 검출 알고리즘 27
• 4.1. 1차 분류기 27
• 4.2. 2차 분류기 31
• 4.3. 3차 분류기 33
• V. 차선 검출 알고리즘 36
• 5.1. 알고리즘 시스템 36
• 5.2. 차선 후보군 생성 38
• 5.3. 차선 검증 및 검출 40
• VI. 실험 43
• 6.1. 실험 환경 및 방법 43
• 6.2. 과속방지턱 검출 결과 46
• 6.3. 차선 검출 결과 50
• VII. 결론 52
• 참고 문헌 53