基于SIFT特征角点匹配算法研究

The Study of the Corner Matching Algorithm Based on SIFT Eigenvector

目标检测是计算机视觉领域中的基础课题之一，它在现代工业、国防和空间技术等领域有着广阔的应用前景。文中阐述了以SIFT特征为基础的角点匹配算法。算法首先介绍了SIFT特征提取步骤，接着以KDTI^e的结构完成特征向量的排序，使用BBF搜索策略完成特征的快速匹配，并以Ransac算法实现精匹配，接着使用最小二乘法完成目标姿态参数的优化，最终完成了目标检测。实验结果表明，算法能够实现对目标发生平移、旋转、尺度缩放、亮度变化和局部遮挡等一系列变化情况下的检测定位，是有效可行的。

Object detection is one of the foundation topics in computer vision. It has broad application prospect in modem industry, national defense and spatial technology. The comer matching algorithm based on SIFT eigenvector is expounded in the article. The algorithm first introduces the processes of the feature distill, detects the sorts the features by KDTree structure, then complete the feature matching by using BBF strategy. It removes the outer points by using Ransac algorithms, op- timizes the localization parameters by using the least squares algorithm, and then locates the interest- ed objects. The experimental results indicate that the algorithm is invariant to image scale and rota- tion, move and part shelter, and change in illumination, and the algorithm is effective and feasible.