基于区域分块与尺度不变特征变换的图像拼接算法

Image mosaic algorithm based on area blocking and SIFT

针对图像匹配算法计算量大,实时性差的问题,提出了一种基于区域分块与尺度不变特征变换(SIFT)相结合的图像拼接算法。该算法利用图像能量的归一化互相关系数快速分割出匹配图像与待匹配图像间的相似区域,利用SIFT算法在重叠区域中搜索出能用于匹配的图像特征点并实现快速精确配准。然后,通过对图像进行了几何校正和图像融合来实现图像序列间的无缝拼接。实验结果表明,该算法减少了传统SIFT算法的大量无用搜索,改善了图像的几何失真,降低了算法复杂度,提高了图像匹配的速度,在保证90%以上的匹配准确率的基础上,计算时间较传统SIFT算法减少了近50%。提出的算法可准确、快速地实现有形变和尺度变换图像的无缝拼接。

For a large amount of calculation and poor real-time performance of image matching algorithms,this paper proposes a novel image mosaic algorithm based on area blocking and Scale Invariable Feature Transformation（SIFT）.This algorithm uses a normalized cross-correlation coefficient from the image energy to search the similar region between the reference image and the image to be matched,and then it takes the SIFT algorithm to extract image feature points in a overlapped region to implement the rapid and accurate image registration.Finally,it performs the image geometry adjustment and image fusion to get a seamless image.The experimental results show that the algorithm reduces the large number of useless search of the traditional SIFT algorithm,ovecomes the image geometric distortion,lowers algorithm complexity,and improves the speed of image matching.With ensuring the accuracy of image matching over 90%,its computation time has decreased by nearly 50% than that of the traditional SIFT algorithm.In conclusion,the algorithm implements accurately and rapidly seamless mosaicking for the images with deformation and scale transform.