SIFT、ORB及KAZE算法的分析和比较

Analysis and Comparison of SIFT,ORB and KAZE Algorithm

特征点提取和匹配是影像处理与分析的基础,目前主要是利用各种算法对影像进行特征点提取与描述。介绍了SIFT、ORB及KAZE算法的原理及步骤,并利用这3种算法对影像进行特征点提取及匹配实验,从运行时间和匹配率上定量地分析这3种算法的优异性。实验结果表明:在影像特征点提取实验,ORB算法提取特征点的速度最快,所用的时间最少,SIFT、KAZE算法其次;在影像特征点匹配实验,KAZE算法的匹配率最高,达39.00%,SIFT、ORB其次。

Feature point extraction and matching are the basis of image processing and analysis. At present, we use various algorithms to extract and describe feature points. This paper introduces the principles and steps of SIFT, ORB and KAZE algorithms, and uses these three algorithms to extract and match the feature points of the images, and analyze the superiority of the three algorithms quanti- tatively from the run time and the matching rate. The results show that the algorithm is based on the feature extraction algorithm, and the algorithm is the fastest and the time taken by the ORB algo- rithm. SIFT and KAZE are the least. In the image feature matching experiment, the matching rate of KAZE algorithm is the highest,reaching 39.00% SIFT,ORB second.