基于四叉树的改进BRISK特征提取算法

The Improved BRISK Feature Extraction Algorithm Based on Quadtree

针对BRISK算法计算速度稍慢、提取的特征点容易出现扎堆的问题,利用四叉树均匀化特征点的方法,提出了基于四叉树的改进BRISK特征提取算法(Quad-BRISK算法):在生成的图像金字塔上提取并检测出具有尺度不变性的特征点之后,采用四叉树方法划分特征点,再计算特征点的方向和BRISK描述子,经过粗匹配、筛选、提纯后最终得到精匹配图像.利用Mikolajczyk和Schmid的特征对比实验图集,对SIFT、ORB、BRISK与Quad-BRISK算法进行了测试对比实验.实验结果表明:Quad-BRISK算法不仅能够提取更加稳定的特征点,同时提高了特征点的匹配精度和计算速度.

In order to solve the problem of low calculation speed of the Binary Robust Invariant Scalable Keypoints(BRISK)algorithm and extreme density of the feature points extracted,an improved BRISK feature extraction algo-rithm based on quadtree(Quad-BRISK algorithm)with the method of quadtree homogenization is proposed.After the feature points with scale invariance are extracted and detected on the generated image pyramid,the feature points are divided with the quadtree method,and then the direction of the feature points and the BRISK descriptor are calculated.After coarse matching,filtering and purification,the precise matching images are finally obtained.Using the feature comparison test dataset of Mikolajczyk and Schmid,the SIFT,ORB,BRISK and Quad-BRISK algorithms are tested and compared.The experimental results show that the Quad-BRISK algorithm can not only ex-tract more stable feature points but also improve the matching accuracy and calculation speed of feature points.