一种全自动稳健的图像拼接融合算 提出了一种全自动稳健的图像拼接融合算法。此算法采用Harris角检测算子进行特征点提取,使提取的

精度达到了亚像素级,然后以特征点邻域灰度互相关法进行特征点匹配得到了初步的伪匹配集合,并运用稳健的

RANSAC算法将伪匹配点集合划分为内点和外点,在内点域上运用LM优化算法精确地估计出了图像间的点变

换关系,最后采用颜色插值对交接处进行颜色过渡。整个算法自动完成,它对有较大误差或错误的特征点数据迭代

过滤,并用提纯后的数据来做模型估计

-A robust fully automatic image mosaic fusion operator presents a fully automatic image stitching robust fusion algorithm. This algorithm uses the Harris operator angle detection feature point extraction, so that the accuracy of extracting the sub-pixel, and then feature points to the neighborhood gray-scale cross-correlation method for matching feature points have been the initial pseudo-match collection and use of sound RANSAC algorithm pseudo-matching point set is divided into inner and outer points, including point-domain LM optimization algorithm used to estimate accurately the image transform relations between points, and finally the use of color interpolation of the color transition of the junction. The entire algorithm for auto-complete, it has a larger error or error of feature points iterative data filtering, and purification of model data to make estimates