论文部分内容阅读
本文针对传统的低空无人机影像拼接处理速度慢和几何精度低的问题,提出了一种快速的无人机影像无缝拼接方法:对输入的原始影像按一定尺度进行降采样,在降采样的影像上进行SURF特征提取和匹配,利用RANSAC方法估计初始的相对单应矩阵,然后用Levenberg-Marquardt方法精化单应矩阵,计算出初始的绝对单应矩阵后利用稀疏光束法平差估计出精确的单应矩阵,通过降采样影像与原始影像的单应关系传递单应矩阵到原始影像级,最后进行影像合成,形成整体拼接图。实验结果表明该方法可以有效地提高拼接速度,解决拼接错位问题。
In this paper, aiming at the problem of low processing speed and low geometric accuracy of traditional low altitude UAV image stitching, a fast UAV seamless stitching method is proposed: down-sampling the input original image according to a certain scale, SURF feature extraction and matching, using the RANSAC method to estimate the initial relative homography, and then using the Levenberg-Marquardt method to refine the homography matrix, the initial absolute homography matrices are calculated and then adjusted using the sparse beam method An accurate homography matrix transmits the homography matrix to the original image level through the homography of the down-sampled image and the original image. Finally, the image synthesis is performed to form an overall mosaic image. Experimental results show that this method can effectively improve the speed of splicing and solve the splicing misalignment problem.