摘要
在使用全天空成像仪对天空进行观测时,由于该设备结构的特点导致采集的云图信息缺失和图形球面畸变,使全天空观测产生误差,为此提出一种全天空云图图像复原算法.针对地基云图的噪声和畸变特点,利用双线性插值收缩图像修复算法对地基云图中固定区域噪声进行复原;然后提出云图遮光带运动预测插值算法和改进型统计扇形插值云图畸变矫正算法,对云图中的遮光带和球面畸变进行图像复原,最终完成对地基云图的修复.实验结果表明,该算法具有普适性,可以更好地复原云图所携带的天空信息.
Information loss and spherical aberrations are main problems in the utilization of total-sky imager cloud image.In this paper,an efficient image inpainting downscaling algorithm based on bilinear interpolation is proposed for noise recovery in fixed areas of cloud images.In addition,a cloud shading motion forecasting algorithm and an improved statistical fan interpolated contours distortion correction algorithm are investigated for the restoration of cloud shading bands and spherical aberration.The combination of the proposed methods can produce reasonable foundation cloud image restoration and information recovery.Experimental results show that the algorithm is universal and can better recover cloud image information.
出处
《计算机辅助设计与图形学学报》
EI
CSCD
北大核心
2014年第6期896-902,共7页
Journal of Computer-Aided Design & Computer Graphics
基金
国家"八六三"高技术研究发展计划(2011AA05A104)
国家电网科技项目(SG11021)
中国电力科学研究院科技创新基金(YN83-14-006)
关键词
云图修复
全天空成像仪
畸变校正
云运动预测
cloud image restoration
total-sky imager
distortion correction
cloud movement prediction
