| 제목 |
감시 카메라 영상의 비조화비를 이용한 차량 사고 분석 |
| 분야 |
ITS/차량 소프트웨어 |
| 언어 |
Korean |
| 저자 |
최영수(대전과학수사연구소), 김종혁(국립과학수사연구원), 윤용문(국립과학수사연구원), 최지훈(국립과학수사연구원), 박종찬(국립과학수사연구원) |
| Key Words |
Vehicle speed(차량 속도), Moving distance(이동 거리), Cross-ratio(비조화비), Digital image processing(디지털 영상처리), Surveillance camera (감시 카메라) |
| 초록 |
Vehicle speed estimation is the most important task in car accident reconstruction and intelligent transportation system design. Recent studies have been conducted in order to evaluate the speed of a vehicle by using the images of various surveillance cameras. However, the limitation is that it is necessary to know whether the calibration coefficient of the camera or only the average speed analysis of a specific interval is possible. In this study, we propose a novel technique for estimating vehicle speed by using a vehicle or structure already recognized as geometric information from the uncalibrated surveillance camera. The digital image processing technique tracks the specific coordinates information in all video frames and calculates the velocity change according to the video frame rate by using a cross ratio, which is the perspective invariant. The results of this method are verified through simulation and actual car evaluation. As a result, it is confirmed that it is possible to accurately analyze the change in the moving speed of a vehicle in different conditions. |
| 원문(PDF) |
다운로드
Journal Site
|
