基于图像纹理特征的SIFT算法研究

针对传统尺度不变特征转换(SIFT)算法存在大量冗余的特征点, 而导致图像匹配过程中运算量大、效率低的问题, 提出一种基于图像纹理特征的SIFT算法。该算法首先采用排列组合熵方法提取图像纹理信息, 在此基础上利用SIFT算法提取特征点, 这样能够减少冗余特征点, 以有效提高算法匹配效率。实验测试结果表明, 该算法与传统的SIFT算法相比, 冗余特征点少, 特征点匹配效率提高到98.04%。

Improved Scale-invariant Feature Transform Algorithm Based on Image Texture

Aiming at the problem of lots of redundancy feature points existing in traditional scale-invariant feature transform(SIFT) algorithm which can led to large amount of computation and low efficiency in the image matching process, an improved SIFT algorithm based on image texture features is proposed. First, the algorithm uses permutation entropy extraction image texture information, on this basis, it uses SIFT to extract feature point, then it can reduce part of the redundancy feature points, in order to effectively improve the efficiency of matching algorithm. The simulation results show that the algorithm has fewer redundant feature points than the traditional SIFT algorithm. And the efficiency of matching feature points increases to 98.04%.