针对航摄影像中存在高亮区域和大面积阴影等灰度剧烈变化区对匀光影像造成的灰度失真现象，本文对Mask匀光算法中的背景影像生成方法进行了改进，通过使用扩张吞噬算法明显改善了背景影像的质量。经对带有高亮度屋顶、高亮度水体和大面积阴影等特殊区域的航摄影像匀光实验表明，本文方法较现行Mask匀光方法具有更好的匀光效果，不但有效消除了特殊区域交界处影像的灰度失真现象，而且明显增强了暗区影像的纹理清晰度。

Dramatic change of the grayscale in aerial images, such as high-brightness region and large areas of shadow, leads to a grayscale distortion phenomenon in dodging images. This study improves the method of background image generation in the Mask dodging algorithm and enhances the quality of background image by using an expansion and engulfment algorithm. The dodging experiment on an aerial image, which includes high-brightness rooftop, high-brightness water, and large areas of shadow, shows that the improved algorithm obtains better result than other current Mask dodging algorithms. The approach eliminates the grayscale distortion problem among the high-brightness and shadow regions effectively and enhances the texture details among the dark regions apparently.