摘要
为解决尺度不变特征转换匹配算法运行时间过长、匹配率不高的问题,提出一种改进的尺度不变特征匹配算法。在原经典的尺度不变特征转换匹配算法的基础上,引入二维Mallat快速小波变换算法,重建图像的低频成分;对高斯金字塔组数进行调整,减少降采样次数;通过优化的随机抽样一致算法剔除误匹配点。MATLAB仿真结果表明,改进后的算法减少了匹配耗时,提高了匹配率,优于原算法。
To solve the problems of scale-invariant feature transform algorithm including the long running time and the low match rate, an improved SIFT algorithm was proposed. On the basis of the original classic scale-invariant feature transform algorithm, a two-dimensional Mallat fast wavelet transform algorithm was introduced, low-frequency components of the image were recon- structed. The number of the Gaussian pyramid group was revised, and the number of down-sampling was reduced. The mis- matching points were finally removed using improved random-sample consensus algorithm. MATLAB simulation results show that the improved algorithm not only reduces the matching time costs, but also improves the match rate, which is better than the original algorithm.
出处
《计算机工程与设计》
北大核心
2015年第8期2129-2132,2142,共5页
Computer Engineering and Design
基金
江苏省自然科学基金项目(BK20131107)
关键词
尺度不变特征转换
小波变换
高斯金字塔
降采样
随机抽样一致
scale-invariant feature transform
wavelet transform
Gaussian pyramid
down-sampling
random-sample consensus
