摘要
为了有效去除薄云对高分光学影像造成的干扰,解决传统算法自动化程度低、去薄云效果不理想的问题,提出改进的薄云最优化变换(Improved Haze-Optimized Transformation,IHOT)算法。首先采用暗原色先验知识从薄云影像中自动选取晴空区,运用薄云最优化变换检测薄云,再创新性地利用植被区域云检测精度较高的特点改进检测结果,最终使用虚拟云点法进行薄云去除。利用高分一号影像进行实验,证明该文提出的算法能够有效地去除薄云对高分影像的影响,尤其对人造地物的色彩和纹理信息恢复效果优于传统算法。
A new algorithm is developed to remove the haze from high-resolution image effectively and improve the automatic de- gree of cloud detection methods. The clear region of haze image is selected automatically based on the prior of dark channel, and the thin cloud detection is generated by improved haze-optimized transformation （IHOT）, then the haze is removed with the vir- tual cloud point method, Using the GF-1 image to make experiment, the proposed algorithm is proved to be better than the tradi- tional algorithms in recoverin~ the color and texture information of artificial objects.
出处
《地理与地理信息科学》
CSCD
北大核心
2015年第1期41-44,F0003,共5页
Geography and Geo-Information Science
基金
民用航天十二五预研项目
关键词
高分影像
去薄云
自动检测
IHOT
high-resolution image
haze removal
automatic detection
IHOT
