CAJ | 학술논문

雾天场景下摄像头采集到的车牌图像质量较低,在车牌识别系统中对图像预处理产生较大影响,从而造成车牌定位和识别效果较差。以雾天环境下实际采集的车牌图像为例,分别采用带色彩恢复的多尺度Retinex算法和暗原色先验去雾算法对实际图像增强进行仿真实验。实验结果表明,利用暗原色先验去雾算法对图像进行增强处理,能够提高图像的对比度,实现去雾效果,从而使车牌定位和车牌识别达到更好的效果。
무천장경하섭상두채집도적차패도상질량교저,재차패식별계통중대도상예처리산생교대영향,종이조성차패정위화식별효과교차。이무천배경하실제채집적차패도상위례,분별채용대색채회복적다척도Retinex산법화암원색선험거무산법대실제도상증강진행방진실험。실험결과표명,이용암원색선험거무산법대도상진행증강처리,능구제고도상적대비도,실현거무효과,종이사차패정위화차패식별체도경호적효과。
The license plate image quality collected by camera in fog scenes is lower,which have a greater impact on image preprocessing in the license plate recognition system,resulting in being less effective in license plate location and identification.A simulation experiment was performed by using color restoration of multi-scale Retinex algorithm and image dehazing algorithms of dark channel priority to enhance actual image based on actual license plate image acquisition in fog.Experimental results show that this method can improve the contrast of the image and filters the fog well so that the better results are achieved in license plate location and recognition.