A method for recognizing an adjacent driving environment based on an SVM raw image is provided. The method comprises the following steps of: obtaining an image taken through an SVM camera; semantically segmenting pixels corresponding to a vehicle (hereinafter referred to as vehicle pixels) and pixels corresponding to a lane (hereinafter referred to as lane pixels) from the image; extracting object information including information on an adjacent lane and information on an adjacent vehicle from the image by post-processing each of the semantically segmented vehicle pixels and lane pixels; and converting the extracted object information into a physical location and transmitting the same to a driving controller. According to the present invention, the SVM raw image is used to recognize the adjacent driving environment to overcome limitations of top-view image processing.
SVM 원 영상을 기반으로 근접 주행 환경을 인식하는 방법이 제공된다. 상기 방법은 SVM 카메라를 통해 촬영한 영상을 획득하는 단계; 상기 영상으로부터 차량에 상응하는 화소(이하, 차량 화소)와 차선에 상응하는 화소(이하, 차선 화소)를 각각 의미 분할하는 단계; 상기 의미 분할된 각 차량 화소와 차선 화소를 후처리하여 상기 영상으로부터 근접 차선 정보와 근접 차량 정보를 포함하는 객체 정보를 추출하는 단계; 및 상기 추출된 객체 정보를 물리적 위치로 변환하여 주행 제어기로 전달하는 단계를 포함한다.
SVM METHOD AND SYSTEM FOR RECOGNIZING A DRIVING ENVIROMENT IN PROXIMITY BASED ON THE SVM ORIGINAL IMAGE
SVM 원 영상을 기반으로 근접 주행 환경을 인식하는 방법 및 시스템
2022-10-27
Patent
Elektronische Ressource
Koreanisch
IPC: | B60W CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION , Gemeinsame Steuerung oder Regelung von Fahrzeug-Unteraggregaten verschiedenen Typs oder verschiedener Funktion |
APPARATUS AND METHOD FOR RECOGNIZING DRIVING ENVIROMENT FOR AUTONOMOUS VEHICLE
Europäisches Patentamt | 2016
|APPARATUS AND METHOD FOR RECOGNIZING DRIVING ENVIROMENT FOR AUTONOMOUS VEHICLE
Europäisches Patentamt | 2016
Programming Enviroment of Control Algorithms
IEEE | 2017
|