基于直方图均衡化的PCA-SIFT图像特征提取与匹配改进算法

Improved Feature Detection and Matching Algorithm for PCA-SIFT Image Based on Histogram Equalization

为了克服SIFT算法运算量大、运算耗时长和实时性不强的缺点,课题组提出了基于直方图均衡化PCA-SIFT算法,以增强灰度图像的明暗对比度,增加匹配成功率,改进描述子生成方式,沿用主成分分析法(PCA)降低特征检测维度,减少耗时。仿真实验表明该算法有效地减少了运算时间,并且一定程度上减少了不必要的特征点的匹配数量。新算法能有效减少运算量。

The SIFT feature extraction algorithm has large computation,the operation takes a long time,and the real-time performance is not strong. In order to enhance the contrast of gray images,increase the success rate of matching,improve the way of descriptor generation,and reduce the dimension of feature detection by using principal component analysis ( PCA),the research group proposed a PCA-SIFT ( principal components analysis-SIFT) algorithm based on histogram equalization. The simulation results show that the algorithm effectively reduces the operation time and to some extent reduces the number of unnecessary feature points matching. The new algorithm can effectively reduce the computational complexity.