基于小波变换多尺度Harris角点检测算法

A wavelet-based multi-scale Harris corner detection algorithm

提出一种新的基于小波变换的Harris多尺度角点检测算法,可以在不同的尺度下获取角点,克服了单一尺度的Harris角点检测算法可能存在的角点信息丢失和易受噪声影响而检测出伪角点等缺点,测试实验表明,与传统Harris角点检测方法比较,本方法有角点检测率高,不易受噪声影响,检测到的角点具有较高的重复率等优点。

A novel wavelet-based multi-scale Harris corner detection algorithm is proposed. It can detect image feature points in different scales and overcome the drawbacks of the traditional Harris corner detector, which often miss significant corners or extract false corners due to noise factor. The experimental results show that the proposed algorithm has a high corner detection rate and a good anti-noise performance. Moreover, comparing with the traditional algorithms, it has a higher repeatability of detected feature points.