摘要
研究航拍图像的拼接问题,提高图像拼接的准确度。由于当通过航拍获取的图像中部分图像相互存在重叠区域的比例不大时,造成拼接不准确。传统的区域的图像拼接算法无法将具有较小重叠区域的图像准确拼接。为了提高航拍图像拼接的准确率,提出一种全局的图像拼接算法,通过使用SIFT算法提取图像的SIFT特征点,根据位置误差最小的原则完成两幅图像的SIFT特征点匹配,最后利用整体最优化方法对拼接结果进行优化,采用全局特征点的拼接方法,可避免传统方法只利用重叠区域灰度特征而不能准确拼接低重叠度图像的问题。实验证明,改进方法利用图像的全局信息,准确地实现图像的拼接,取得了满意的结果。
Research mosaic problem of aerial images to improve the accuracy of the image mosaic.When the overlapping part proportion in aerial images is not large,the traditional local image mosaicing algorithm can not implement the image mosaic effectively with smaller overlap areas in the aerial images,resulting in the low accuracy of aerial image mosaicing.This paper proposed an algorithm based on global vision.Through using the SIFT algorithm to extract SIFT feature points in the image,according to the principle of the minimum position error,the SIFT feature points matching of two images was finished.Then using the whole optimized method to optimize the finished mosaic image.Experiments show that the method can use the information of the whole image and accurately realize image mosaic.
出处
《计算机仿真》
CSCD
北大核心
2012年第4期282-285,300,共5页
Computer Simulation
关键词
航拍图像
图像拼接
全局信息
Aerial images
Image mosaic
Global information
