无人机遥感图像融合方法研究

针对待拼接的无人机遥感图像重叠区域不规则、焦距不固定、含噪声较多等问题, 在加权平均融合算法的基础上引入基于折线化思想的不规则重叠区域生成法, 减小算法误差, 并运用基于距离比的自适应算法实现权值自动匹配, 使算法在不受重叠区域形状限制的同时得到更精确的计算结果, 解决了图像拼接后融合区域分辨率低、拼接线明显的问题。仿真表明, 改进的加权平均融合算法在保持原算法快速性的同时, 达到了去除拼接缝隙、使图像融合区域过渡平滑的目的, 并获得了融合质量较好的大视野无缝拼接图像。

Research on Remote Sensing Image of UAV Fusion Algorithm

The solution to the problems of low resolution and obvious stitching line in fusion area after image mosaic is proposed, by introducing irregular overlapping region generating method based on piecewise linear approximation to reduce algorithm error and using adaptive algorithm based on distance ratio to implement the automatic matching of weights, which can make sure the algorithm get more accurate results without being restricted to the shape of overlapping region. The simulation showed the improved weighted average fusion algorithm achieved the purpose of removing the stitching gap and smoothing the fusion area with the same rapidity as original algorithm's and was able to obtain large field of view images with better fusion quality.