摘要
针对传统SIFT匹配算法数据量大、时间复杂度高的问题,提出基于尺度不变特征变换(SIFT)特征提取方法获得特征点,并采用变换步长的圆形区域选区对特征点进行描述,改进了SIFT特征的64维描述符和88维描述符的不足。将改进后的算法应用到图像拼接过程中,通过实验验证了改进后的方法在时间复杂度方面有所改善。
Since the traditional scale invariant and feature transform （SIFT） algorithm has large data size and high time complexity,a feature extraction method based on SIFT is proposed to acquire the feature points. The region in circular area with step？conversion is adopted to describe the feature points,and the insufficient of 64？dimensional descriptor and 88？dimensional descriptor of SIFT feature is improved. The improved algorithm was applied to image mosaic process. The experiment results veri？fy that this method has great improvement in the aspect of time complexity.
出处
《现代电子技术》
北大核心
2015年第19期72-75,78,共5页
Modern Electronics Technique
