基于尺度不变特征的目标识别方法

Target Recognition Method Based on Scale Invariant Feature

针对普通图像中的目标识别问题,在SIFT算法的基础上,利用特征点的主方向信息对SIFT算法进行改进,提出一种适应性强、识别准确率高的目标识别方法。实验结果表明,在SIFT特征匹配之前剔除主方向差异较大的特征点对,不仅减少了特征匹配的运算量,还提高了执行效率和识别准确率。

Aiming at the problem of target recognition in common images,based on SIFT algorithm,the SIFT algorithm is improved by using the main direction information of feature points,and a target recognition method with strong adaptability and high recognition accuracy is proposed.The experimental results show that eliminating the feature point pairs with large difference in main direction before SIFT feature matching can not only reduce the computation amount of feature matching,but also improve the execution efficiency and recognition accuracy.