| 摘要: |
| 同时检测出包括边缘特征和线特征在内的多种图像特征一直是图像特征检测中的难点。在相位一致性基本原理的基础上，提出了一种基于局部频率的相位一致性检测图像特征的算法。试验结果表明，采用的算法在图像特征的细节保留和定位方面要优于传统的相位一致性算法和Canny算法。 |
| 关键词: 图像处理 特征检测 相位一致性 局部频率 相位偏移 |
| DOI：10.3969/j.issn.1001-893X. |
| 投稿时间：2006-01-14修订日期：2006-04-06 |
| 基金项目: |
|
| An Improved Algorithm of Phase Congruency in Image Features Detection |
|
| () |
| Abstract: |
| In image features detection it is difficult to get a wide range of interesting features including edges and lines. An algorithm of phase congruency based on localized frequencies is proposed in image features detection on the basis of its fundamental theory. The simulation results show that the proposed method is more efficient than traditional phase congruency and Cannyg method in the detail preserving and location of image features. |
| Key words: image processing feature detection phase congruency localized frequency phase shifting |
《电讯技术》编辑部 