测绘科学技术学报Issue(2)：173-176,4.DOI:10.3969/j.issn.1673-6338.2014.02.015
一种改进的 SIFT 特征提取算法
An Improved SIFT Feature Extraction Algorithm
摘要
Abstract
In order to meet the need of real-time ability in image positioning on UAV, an improved SIFT feature ex-traction algorithm is put forward. Firstly, setting proper pixel step when testing scale space feature points, to re-duce the testing times of extreme points and make feature points uniformly distributed. Then, the contrast threshold is determined based on the gray scale of tested feature points and the points with low contrast are eliminated. The feature extraction can be accelerated by using adaptive pixel steps and contrast thresholds. Experiment results show that the improved algorithm can reduce the numbers of the extracted feature points and improve the algorithm effi-ciency greatly.关键词
SIFT 算法/像素步长/对比度阈值/特征点/算法效率Key words
SIFT algorithm/pixel step/contrast threshold/feature points/algorithm efficiency分类
天文与地球科学引用本文复制引用
牛俊伟,郝向阳,刘松林..一种改进的 SIFT 特征提取算法[J].测绘科学技术学报,2014,(2):173-176,4.
