영상처리를 이용한 운전자 보조 시스템에 관한 연구

Abstract

Recently, there have been a growing interest in self-driving cars development and advanced driver assistance system (ADAS) according to the development of IT technology and automobile control technology. Self-driving cars operate vehicle control system using various kinds of sensors that replace the driver's visual information. A lane departure warning system (LDWS) and a forward collision warning system (FCWS) can be realized by processing image obtained only through the vehicle front camera. This thesis describes three contents for acquiring images through the vehicle front camera and implementing the driver assistance system through image processing. The first describes related researches to obtain information to be used in the LDWS and the FCWS. The second describes the LDWS through the lane detection, and finally the FCWS through vehicle detection will be described. To evaluate the performance of the proposed algorithm, this thesis measures the detection ratio of valid noises such as a crosswalk, a stop line, etc. and evaluates the valid warning of the LDWS through comparing with the vehicle side image. The recognition rate for noise was 97.6% and the average accuracy was 96.8% in the valid warning of the LDWS. Also, the valid warning of the FCWS is evaluated by comparing the actual motion with the detection result and the average detection rate was 96.2%.