论文部分内容阅读
提出了一种利用零交叉点特征提取的改进SIFT算子用于遥感影像的自动匹配。将图像几何特征引入到尺度空间探测中,获得了重复性更高、更稳定的特征。采用近景数码大旋角数码立体像对和低空航摄立体像对进行了算法测试。实验表明,改进后的SIFT算子应用于遥感影像自动匹配,在特征提取重复率、匹配正确点数、匹配正确率上均有明显提升。
An improved SIFT operator using zero-crossing feature extraction is proposed for automatic matching of remote sensing images. The introduction of the geometric features of the image into the scale space exploration has achieved the characteristics of higher repeatability and stability. The algorithm is tested by using near-digital digital large-angle digital stereo pair and low-altitude aerial photo-stereo pair. Experiments show that the improved SIFT operator is applied to the automatic matching of remote sensing images, which improves obviously in feature extraction repetition rate, matching correct points and matching accuracy.