清远职业技术学院学报
清遠職業技術學院學報
청원직업기술학원학보
JOURNAL OF QINGYUAN POLYTECHNIC
2011年
6期
73-76
,共4页
车道偏离预警系统%车道识别%Sobel算子%感兴趣区域分割
車道偏離預警繫統%車道識彆%Sobel算子%感興趣區域分割
차도편리예경계통%차도식별%Sobel산자%감흥취구역분할
Lane offset warning system%Lane detection%Sobel operator%Region of interest segmentation
为了准确识别出高速公路图像中车道的标志边缘,结合道路图像特点,首先采用Sobel边缘算子对中值滤波后的道路图像进行增强,进而用Otsu算法将图像分割,在感兴趣区域运用基于像素的运算方法去除噪声,最后利用Hough变换进行拟合,识别出车道线。结果表明,对于2616×3488像元的道路图像,特征点与拟合直线的最大偏差为3.7263像元,能够准确识别出图像中的车道线,为智能汽车的安全预警提供了有益的探索。
為瞭準確識彆齣高速公路圖像中車道的標誌邊緣,結閤道路圖像特點,首先採用Sobel邊緣算子對中值濾波後的道路圖像進行增彊,進而用Otsu算法將圖像分割,在感興趣區域運用基于像素的運算方法去除譟聲,最後利用Hough變換進行擬閤,識彆齣車道線。結果錶明,對于2616×3488像元的道路圖像,特徵點與擬閤直線的最大偏差為3.7263像元,能夠準確識彆齣圖像中的車道線,為智能汽車的安全預警提供瞭有益的探索。
위료준학식별출고속공로도상중차도적표지변연,결합도로도상특점,수선채용Sobel변연산자대중치려파후적도로도상진행증강,진이용Otsu산법장도상분할,재감흥취구역운용기우상소적운산방법거제조성,최후이용Hough변환진행의합,식별출차도선。결과표명,대우2616×3488상원적도로도상,특정점여의합직선적최대편차위3.7263상원,능구준학식별출도상중적차도선,위지능기차적안전예경제공료유익적탐색。
In order to identify the edge in the motorway lane accurately, with the characteristics of road image, the road image which is filtered by median filter is enhanced by the Sobel operators. And then the image is segmented using Otsu algorithm. In the region of interest, the noise is removed using the computational method based on the pixels. At last, the lanes were identified by using Hough transform. The maximum offset between characteristic points and the fitting straight line is 3.7263 pixels. The experimental results are presented showing that the lanes can be identified accurately for 2616 × 3488 road image. Intelligent vehicle safety warning system is provided a useful exploration.