CAJ | 학술논문

针对在雨雾霾天气条件下，大气介质的散射和吸收作用导致光电成像系统接收的图像对比度降低，细节模糊不清及颜色偏移，提出通过快速图像复原来解决此类图像退化问题。基于大气成像光学模型，在暗通道先验的理论基础上，提出了一种基于形态学滤波器的快速估算暗通道图像的方法，并采用参数自适应调整方法来抑制暗通道先验不满足时的大片天空/白墙区域的颜色失真现象。实验结果表明，该算法能够有效快速复原雨雾天气条件下的降质图像，对于600×400大小的图像，其Matlab复原仿真时间仅为0.4 s，复原后的图像主观视觉质量明显提升，其大片天空/白墙区域的颜色失真得到有效抑制。
침대재우무매천기조건하，대기개질적산사화흡수작용도치광전성상계통접수적도상대비도강저，세절모호불청급안색편이，제출통과쾌속도상복원래해결차류도상퇴화문제。기우대기성상광학모형，재암통도선험적이론기출상，제출료일충기우형태학려파기적쾌속고산암통도도상적방법，병채용삼수자괄응조정방법래억제암통도선험불만족시적대편천공/백장구역적안색실진현상。실험결과표명，해산법능구유효쾌속복원우무천기조건하적강질도상，대우600×400대소적도상，기Matlab복원방진시간부위0.4 s，복원후적도상주관시각질량명현제승，기대편천공/백장구역적안색실진득도유효억제。
Under haze , fog, and rain weather conditions , the image received by a photoelectric imaging sys-tem will lose the contrast and color fidelity owing to the absorption and scattering in atmosphere .To overcome the image degradation mentioned above , this paper proposes a new dark channel image estimation method u-sing morphology filter based on the analysis of the atmospheric optical model and the dark channel priority .It further utilizes a dynamic parameter strategy to solve the color distortion in a large area of sky or white wall . Results on a variety of outdoor haze images show that the proposed method can effectively restore a haze -de-graded image and improve the image quality .As for the image with 600 pixel ×400 pixel, the restoration sim-ulation time by Matlab is only 0.4 s.The dynamic parameter strategy is simple but effective to reduce the color distortion phenomenon .