摘要
在研究传统Harris角点检测算法的基础上,结合小波变换的多分辨率特性,提出一种基于小波变换边缘检测与Harris角点检测的多尺度图像配准算法。该算法保持了Harris角点检测算法在图像旋转、缩放或灰度变化时角点提取效果依然良好的优点,改进了其对噪声敏感且不具有尺度不变性的缺点,在不同尺度上有较高稳定性与匹配精准度,抗噪性较强。
Based on the study of traditional Harris comer detection algorithm, combined with the multi-resolution features of wavelet transform, a muti-scale image registration algorithm based on the wavelet edge detection and Harris comer detection is proposed. When the image exist rotation, zooming or gray change, the comer detection effect of Harris algorithm is still good. The algorithm improves the shortcomings of noise sensitive and not having scale invariance, have strong noise immunity, higher stability and matching accuracy in different scales.
出处
《天津职业技术师范大学学报》
2011年第3期36-39,共4页
Journal of Tianjin University of Technology and Education
基金
天津市高等学校科技发展基金项目(20080708)
天津市应用基础及前沿技术研究计划(09JCYBJC27300)
关键词
图像配准
小波变换
角点检测
image registration
wavelet transform
comer detection
