CAJ | 학술논문

车牌定位是车牌识别系统的关键，为了提高复杂光照下车牌定位的效率，提出了一种基于车牌图像预处理和纹理边缘特征相结合的车牌定位方法。在预处理过程中利用HSV颜色空间中亮度和色度相互独立的特性，对图像进行亮度处理，得到了一幅有效去除背景干扰且亮度适度的二值图像；然后采用Canny算子对车牌进行边缘检测，得到大致的车牌区域；最后对车牌分别进行水平和垂直投影，得到精确的车牌区域。整个过程用Mat1ab软件实现，结果证明该方法在复杂光照下具有较强的适应性与稳健性，有一定的实用价值。
차패정위시차패식별계통적관건，위료제고복잡광조하차패정위적효솔，제출료일충기우차패도상예처리화문리변연특정상결합적차패정위방법。재예처리과정중이용HSV안색공간중량도화색도상호독립적특성，대도상진행량도처리，득도료일폭유효거제배경간우차량도괄도적이치도상；연후채용Canny산자대차패진행변연검측，득도대치적차패구역；최후대차패분별진행수평화수직투영，득도정학적차패구역。정개과정용Mat1ab연건실현，결과증명해방법재복잡광조하구유교강적괄응성여은건성，유일정적실용개치。
License p1ate 1ocation is the key part of the 1icense p1ate recognition system. To improve the effi-ciency of 1icense p1ate 1ocation in comp1ex 1ighting conditions, a method based on the pretreatment of 1icense p1ate image and the features of edge and texture was proposed. In the pretreatment process, the 1icense p1ate image was brightened by using the features of independence between brightness and co1or in HSV co1or space, and a binary image with suitab1e brightness was obtained after efficient1y e1iminating the background interference. Then the approximate position of the 1icense p1ate was got by using Canny edge detection to ex-amine the edges. Fina11y, the horizonta1 and vertica1 projections of 1icense p1ate were checked respective1y, and the precise 1ocation of a 1icense p1ate was obtained. The who1e process was rea1ized using Mat1ab soft-ware. The resu1t shows that the method has stronger adaptabi1ity and robustness in comp1ex 1ighting condi-tions, and it cou1d be used wide1y.