摘要
针对无人机遥感影像因重叠度高、数据量大等因素导致影像融合难以兼顾高效性和擦除多度重叠区域的拼接痕迹等问题，提出了一种改进的P范数融合方法。该方法首先将待融合图像按中心距离法计算初始权值；然后根据像素坐标位置自动计算指数，统计初始权值的指数得到最终的权值，进一步扩大了像素权值的相对差异；最后对源影像进行加权融合。实验表明：改进算法不仅能改善无人机影像多度重叠引发的重影，还能有效抑制局部模糊现象。
Since there are high degree of overlap and huge volume of data in UAV remote sensing images, it is difficult to maintain the high efficiency and erase mosaicking edges in image fusion. Therefore, an improved P norm fusion method is pro- posed in this paper. Firstly, the initial weight value of images waiting for fusion is calculated by the center distance method, and then the index is calculated automatically according to pixel coordinates. The final weight value is obtained based on statistics of initial value and the relative difference of weight value of pixel is enlarged in the process. Finally, weight fusion of source image is conducted. The experiment indicates that the improved algorithm will reduce double images caused by muhi-overlap and partial vagueness effectively.
出处
《测绘科学与工程》
2016年第6期41-46,共6页
Geomatics Science and Engineering
关键词
无人机遥感影像
图像融合
改进的P范数法
重影
模糊
UAV remote sensing image
image fusion
improved P norm method
double images
vagueness
