一种卫星多光谱图像亚像元波段配准精度自动评价方法

多光谱图像波段配准精度是多光谱影像数据的一项重要指标。文章实现了一种基于相位一致特征的多光谱图像亚像元波段配准精度自动评价方法。通过对CBERS-02B卫星CCD多光谱图像波段配准精度自动评价,表明该方法对于亮度和对比度具有不变性,能均匀稳定提取多光谱数据不同波段影像大量同名特征匹配点,自动给出影像的不同区域的亚像元配准精度,实现对整景影像数据的较为完整准确评价。根据评价结果进行了地面处理系统校正,使行列方向除局部区域外均达到配准精度。

An Automated Evaluation Method of Sub-pixel Band-to-band Registration Accuracy for Satellite Multispectral Image

Band-to-band registration accuracy is an important parameter of multispectral image data.A high precision method based on the phase congruency features was presented in this paper for automated evaluating sub-pixel band-to-band registration accuracy of satellite multispectral imagery.The sub-pixel band-to-band registration accuracy automatic assessment results of CBERS02B CCD multi-spectral remote sensing imagery data showed that the proposed method was invariant to intensity and contrast variations,and significant features could be detected and matched reliably between different band images,the rule of band-to-band misregistration and the results of band-to-band registration are automatic presented completely.The further band calibration of sub-pixel ground data processing systems was performed based this results and a good registration accuracy in row and line direction had been achieved except some local areas.