基于旋转SURF算子的图像配准新方法

针对SURF算法中快速Hessian矩阵行列式检测出的特征点的不连续现象,从而造成的旋转,模糊和光照变化适应性较差的不足,提出一种旋转SURF检测算子的图像配准新方法。该算法通过将SURF算法的积分图像盒子滤波模板逆时针旋转45度,引入一种可以检测角度旋转的滤波核提升检测算子对不同图像变换的匹配性能,保证新的检测算子与原算法较好的结合,同时利用改进的单纯形算法依据输入图像进行参数优化。仿真结果表明,该方法不仅保留了算法的速度优势,缩短了配准时间,而且在图像模糊变换,光照变换和JPEG压缩变换方面性能有明显的提升,此外对视角变换以及小尺度变换性能也有提高。

A new image registration algorithm based on rotated SURF

Weaknesses in the Fast Hessian detector utilized by the SURF algorithm make it less robust to image rotation, image blurring, illumination change and other transformations. In order to solve this problem, an alternative to the SURF detector is proposed which utilizes filters that are rotated 45 degrees counter-clockwise. This new detector is robust to various image transformation and has the ability to match the original SURF algorithm. The new algorithm is also improved by the simplex algorithm (ISA) which can optimize parameters based on input image. Performance testing shows that the new method retains the speed advantage designed into the original SURF algorithm and outperforms the regular SURF detector when subject to image blurring,illumination changes and JPEG compression. The new method outperforms regular SURF slightly when subjected to affine changes and small image scale transformations.