CAJ | 학술논문

SIFT(尺度不变特征变换)是一种基于多尺度空间的图像局部特征提取算法,可以很好的应用于图像的匹配.基于SIFT的图像匹配包括尺度空间的生成、局部极值点的检测、特征点位置的精确定位、特征点主方向的计算、SIFT描述子的生成和图像匹配方法等步骤,该方法适应性强,匹配效果好.实验表明,SIFT算法对图像中的各种变化具有较强的鲁棒性.
SIFT(척도불변특정변환)시일충기우다척도공간적도상국부특정제취산법,가이흔호적응용우도상적필배.기우SIFT적도상필배포괄척도공간적생성、국부겁치점적검측、특정점위치적정학정위、특정점주방향적계산、SIFT묘술자적생성화도상필배방법등보취,해방법괄응성강,필배효과호.실험표명,SIFT산법대도상중적각충변화구유교강적로봉성.
SIFT is a kind of local image feature extraction algorithm based on multi scale space,which can well applied in image matching.This paper introduced the formation of scale space,detection of the local extreme point,precise positioning of feature points location,calculation of the main direction of the feature point,generate SIFT description and image matching method based on SIFT.Experiments show that SIFT is robust to various changes in the image.