基于灰度校正的SIFT算法全景图实现  

Realization of panoramic images based on scale invariant feature transform

作　　者：艾达[1] 乔明明[1] 高波[2] 

机构地区：[1]西安邮电大学公安部电子信息勘查应用技术重点实验室,陕西西安710121 [2]西安邮电大学电子工程学院,陕西西安710121

出　　处：《西安邮电大学学报》2014年第4期77-81,共5页Journal of Xi’an University of Posts and Telecommunications

摘　　要：提出应用灰度校正改进全景图像无缝拼接的方法。利用直方图均衡化技术消除拍摄图像的灰度不一致;应用基于尺度不变特征变换算法(SIFT)对图像进行特征匹配,最后利用加权平均算法进行图像融合实现图像的无缝拼接。实验结果表明,图像的匹配对增加了86%,匹配效果优于现有SIFT方法。An image grey level correction method is proposed in this paper to achieve the image seamless splicing.Histogram equalization is used to eliminate the inconsistent grey level of the captured image.Captured images are then matched by scale invariant feature transform and the seamless mosaic fused image is created by using linear smoothing algorithm.Experiment results show that the matching pairs are increased by 86%in the image and matching result is better than that of the conventional SIFT method.

关 键 词：柱面全景图 尺度不变变换 特征匹配 图像拼接 灰度校正 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]