基于亮度校正的航拍图像拼接融合算法

Stitching Fusion Algorithm of Aerial Image Based on Brightness Correction

针对无人飞行器航拍图像拼接可能产生拼接缝及拼接过程中占用内存过大的问题,提出一种基于亮度校正的航拍图像拼接融合算法。首先,对配准的航拍图像进行亮度校正预处理;然后,搜索图像获得最佳缝合线,并从图像重叠区域划分出过渡区域,减小拉普拉斯金字塔面积;最后,在不同区域分别采用加权融合算法和拉普拉斯金字塔融合算法对图像进行拼接。实验结果表明,提出的算法能够有效改善曝光差异,同时减小拼接时所占用的内存。

Aiming at the problem that there may be stitching seams,and occupy too much memory in unmanned aerial vehicles(UVA)image stitching,an aerial image stitching fusion algorithm based on brightness correction was proposed.Firstly,the aerial images were registered and preprocessed by brightness correction.Then,the best suture line was obtained by searching the images,and the image overlap area was divided into transition areas,which can reduce the Laplacian pyramid area.Finally,images were stitched in the different areas with weighted fusion algorithm and Laplacian pyramid fusion algorithm,respectively.Experiments show that the proposed algorithm can effectively improve the exposure difference,and at the same time the memory occupied during stitching can be reduced obviously.