CAJ | 학술논문

针对智能车辆安全辅助驾驶系统中利用单目视觉进行车道识别的问题，提出了1种基于平行直线对模型的车道检测方法。该方法根据高速公路图像特征构建平行直线对模型，在此基础上先利用 Hough变换提取直线，再由改进的级联 Hough变换检测出平行直线对的消失点，最后通过消失点和先验信息来提取当前车道线。使用M atlab对高速公路上不同路段、不同光照情况、不同车辆干扰下共150幅道路图像进行实验，检测精度达88．6％，平均检测时间为0．24 s。实验结果表明，这一方法在高速公路行驶环境下能较准确地检测出当前车道线，具有很好的光照适应性、抗车辆干扰性和一定的实时性。
침대지능차량안전보조가사계통중이용단목시각진행차도식별적문제，제출료1충기우평행직선대모형적차도검측방법。해방법근거고속공로도상특정구건평행직선대모형，재차기출상선이용 Hough변환제취직선，재유개진적급련 Hough변환검측출평행직선대적소실점，최후통과소실점화선험신식래제취당전차도선。사용M atlab대고속공로상불동로단、불동광조정황、불동차량간우하공150폭도로도상진행실험，검측정도체88．6％，평균검측시간위0．24 s。실험결과표명，저일방법재고속공로행사배경하능교준학지검측출당전차도선，구유흔호적광조괄응성、항차량간우성화일정적실시성。
Aiming at improving the accuracy of traditional way of using monocular vision to identify the running lane ,a new method of lane detection based on the double straight line model is proposed .According to the characteristics of highways ,the method firstly builds a double straight line model ,and extracts straight lines by Hough Transformation , then detect the vanishing point of parallel lines by an improved Hough Transformation ,and finally extracts current traffic lanes through vanishing point and prior information from an image .An experiment is conducted based on 150 road images from different road sections ,with different illumination levels and different vehicle interferences .It is found out from the study results that the extraction accuracy is 88 .6% and the average detection time is 0 .24 seconds .Experimental results show that the proposed method can detect highway lanes with a good accuracy and it is also good at adapting to different illumination levels ,reducing the impact of vehicle disturbance and processing data in real-time .