CAJ | 학술논문

针对视频流中多个车牌的定位问题，提出了一种基于图像复原技术与HSV颜色空间的多车牌定位方法。该方法的主要思想是用改进的维纳滤波器对获取的视频图像进行复原，根据车牌的颜色特征，利用 HSV 颜色模型构造5级灰度图，对灰度图用形态学方法进行消噪和区域连通处理，最后利用水平、垂直投影技术，结合车牌先验知识，通过分析和判断来确定并分割出各个汽车牌照。实验表明，该方法能有效的从复杂的背景中准确定位并提取出多个运动中的车牌。
침대시빈류중다개차패적정위문제，제출료일충기우도상복원기술여HSV안색공간적다차패정위방법。해방법적주요사상시용개진적유납려파기대획취적시빈도상진행복원，근거차패적안색특정，이용 HSV 안색모형구조5급회도도，대회도도용형태학방법진행소조화구역련통처리，최후이용수평、수직투영기술，결합차패선험지식，통과분석화판단래학정병분할출각개기차패조。실험표명，해방법능유효적종복잡적배경중준학정위병제취출다개운동중적차패。
For the location of multiple License Plates in video streams, this paper presented an effective method of Multi-license plates location based on image restoration and HSV color model. The main idea of this method was to use an improved Wiener filter to get the video image recovery. And then in HSV color space, it constructed a five-scale gray image according to the color characteristics of license plate, and used the morphology smooth to eliminate noise. Finally, the potential region of the plate was segmented through horizontal and vertical projection value and we got the accurate position of the multiple plates by prior knowledge of the car-plate region. Experimental results show that the proposed method can extract multiple motive license plates exactly from a complex background.