CAJ | 학술논문

提出了一种基于颜色空间理论和形态学结合的方法。首先对图像进行颜色空间转换，按要求提取出需要的特征颜色区域，再对颜色特征区域进行检测，对边缘进行灰度统计，按一定的比例进行筛选，并利用前方车辆位置变化的特点，对下一次车牌可能出现的区域进行粗定位，利于下一次更快速的定位。应用该算法对100幅车牌图像进行定位，定位准确率达90％，速度均在0．1s内。
제출료일충기우안색공간이론화형태학결합적방법。수선대도상진행안색공간전환，안요구제취출수요적특정안색구역，재대안색특정구역진행검측，대변연진행회도통계，안일정적비례진행사선，병이용전방차량위치변화적특점，대하일차차패가능출현적구역진행조정위，리우하일차경쾌속적정위。응용해산법대100폭차패도상진행정위，정위준학솔체90％，속도균재0．1s내。
This paper presents a combination of theory and morphology of the color space. Firstly, image color space is conversed, and the required features of the color region detection to is extracted, Then, the color characteristics of the regional are tested, the gray level statistics on the edge is done, and is filtered by a certain percentage. By using characteristics of the changes of the vehicle location on the front, the coarse positioning of the possible license plate region for the next time is finished, which is benefit of a more rapid positioning for the next time. 100 license plate image are used, for positioning in this which is benefit of algorithm, the positioning accuracy rate is 90%, the speed is in 0.1 s.