论文部分内容阅读
中国西南丘陵常态山和喀斯特山交错分布,遥感影像普遍存在山体阴影,分布零散且无规律,基于DEM的地形校正模型(C校正等)虽然算法成熟、易于操作,但在复杂地形区存在误差。引入基于相似像元包络线的阴影校正方法(CR校正),按照阴影提取、包络线去除、相似像元寻找和阴影亮度重建的步骤,采用西南丘陵地区Landsat 8 OLI影像进行验证实验。结果表明:CR校正后,阴影区的视觉特征与邻近非阴影区趋于一致,阴影像元亮度有明显提升;校正后影像主要波段标准差减小,与非阴影区参考光谱的相对均方根误差在2.919%以内,最低仅为0.516%;自动分类精度从43.59%提高到61.57%,CR校正有效提高了有阴影的丘陵地区遥感影像质量。
The normal mountains and the karst hills in southwestern China are staggered. The mountain shadows are scattered widely and irregularly in the remote sensing images. Although the terrain correction model based on DEM (C correction, etc.) is mature and easy to operate, there are errors in the complex terrain. A shading correction method (CR correction) based on similar pixel envelopes is introduced. According to the steps of shading extraction, envelope removal, similar pixel finding and shadow brightness reconstruction, Landsat 8 OLI images in southwestern hilly region are used to verify the experiment. The results show that after CR correction, the visual characteristics of the shaded area tend to be consistent with those of the adjacent non-shaded areas, and the shades of the shaded pixels have been significantly improved. The standard deviation of the main band of the corrected image decreases and the relative root mean square of the non-shaded reference spectrum The error is within 2.919%, the lowest is only 0.516%. The accuracy of automatic classification is improved from 43.59% to 61.57%. The CR correction effectively improves the remote sensing image quality in shaded hilly areas.