CAJ | 학술논문

为提高高速公路收费道口自动车型识别系统（Vehicle Recognition System，VRS）的识别准确率，针对基于三维成像的VRS中的车轮检测，提出一种优化算法，以精确检测出车轮的位置和大小。利用激光雷达检测到有疑似车轮的目标通过扫描平面时，控制摄像头抓拍图像；对抓拍到的图像进行灰度化、图像增强、边缘提取、背景分离等处理；利用霍夫变换进行车轮边缘拟合，检测图像中是否含有车轮，若有，可同时得到车轮的位置和大小；在MATLAB中实现该算法。结果表明，该车轮检测算法能够快速、精确地检测出车轮，效果良好。
위제고고속공로수비도구자동차형식별계통（Vehicle Recognition System，VRS）적식별준학솔，침대기우삼유성상적VRS중적차륜검측，제출일충우화산법，이정학검측출차륜적위치화대소。이용격광뢰체검측도유의사차륜적목표통과소묘평면시，공제섭상두조박도상；대조박도적도상진행회도화、도상증강、변연제취、배경분리등처리；이용곽부변환진행차륜변연의합，검측도상중시부함유차륜，약유，가동시득도차륜적위치화대소；재MATLAB중실현해산법。결과표명，해차륜검측산법능구쾌속、정학지검측출차륜，효과량호。
In order to improve the recognition accuracy rate of automatic Vehicle Recognition System (VRS)for highway toll lanes,in view of the wheel detection in three-dimensional imaging-based VRS, an optimized wheel detection algorithm is proposed. This algorithm can accurately detect the position and size of a wheel. When an object suspected to be a wheel is detected passing the scanning plane by laser radar,a camera is controlled to capture the image. The image is processed by graying,image enhancing, edge extracting,and background separating. The wheel edge is fitted by Hough transform to detect whether there is a wheel in the image. If so,the position and size of the wheel can be obtained at the same time. The algorithm is implemented in MATLAB. The result shows that the algorithm can detect wheels quickly and accurately and is of good effect.