一种有效抑制机载传感器成像畸变干扰的景象匹配算法

A scene matching algorithm for effective imaging distortion disturbance restrain of airborne sensor

电荷耦合器件(CCD)与合成孔径雷达(SAR)是两种最常用的机载传感器,是景象匹配辅助导航系统的关键组成部分。当系统传感器获取的实测图存在严重成像畸变时,传统的基于特征点集间Hausdorff距离测度的匹配算法精度较低,通过引入多尺度自卷积(MSA)这种具有仿射不变性的局部特征,并将其同图像的尺度不变变换(SIFT)特征串行组合可以得到一种更加鲁棒的匹配算法。采用CCD和SAR传感器获取的图像对算法进行了仿真验证,实验结果表明:引入MSA特征对抑制景象匹配算法中多源传感器成像畸变的干扰十分有效,匹配精度显著提高。

Charge coupled device（ CCD） and synthetic aperture radar（ SAR） are two typical airborne sensors and key component of scene matching aided navigation system. When sensing image is with considerable distortion,traditional matching algorithm based on Hausdorff distance feature points has very low precision,by introducing multi-scale auto-convolution（ MSA）,which has affine invariant feature,is combined with scale invariant feature transform（ SIFT） features by serial combination strategy to form a more robust matching algorithm. Simulation experiments is carried out on acquired by images CCD and SAR sensor,the result shows the proposed algorithm can efficiently resist the distortion of sensing image and matching precision is improved obviously.