GSSAC:一种用于遥感影像配准的误匹配点检测方法

GSSAC: false matching points detection method for remote sensing images

遥感影像配准中,由于光照、成像角度、几何变形等因素的影响,无论采用何种配准方法,总会产生误匹配点,因此误匹配点检测也是一个非常重要的步骤。针对常用RANSAC(random sample consensus)方法不稳定、无法准确检测分布不均匀匹配点的缺点,提出了分组排序采样一致性(group sorted cample consensus,GSSAC)方法来提高误匹配点检测的稳定性和精度。分组排序采样方法首先将匹配点分为若干组,在每组内计算匹配点的误差并排序,然后在每组中分别采样若干个匹配点组成估算模型参数需要的匹配点。实验结果表明,GSSAC方法可以稳定地获得高精度的检测结果。

Given the influences of illumination, imaging angle and geometric distortion, among others, false matching points still occur in all image matching algorithms. Therefore, false matching points detection is an important step in remote sensing image registration. RANSAC was typically used to detection. However, RANSAC couldn＇ t accurately detect unevenly distribu- ted matching points and the result was unstable. To overcome these disadvantages and improve accuracy and stability, this paper proposed GSSAC method. Group sorted sample strategy first divided all matching points into several groups. Then,it sorted the matching points in each group. Finally, one or more matching points were sampled based on their orders for each group. These matching points were formed by the minimum number of points required to compute model parameters. The experiments show that the stability and accuracy of GSSAC are better compared with other methods.