针对航空面阵摆扫数据无控制点几何校正问题，提出了一种基于POS数据的面阵摆扫热红外影像的分步几何校正方法，主要包括利用视矢量法对影像校正、投影至高斯地辅平面进行拼接、及提取虚拟控制点三个步骤。利用最小二乘法对全局影像进行第二次几何精校正，实现研究区热红外图像几何校正。实验图像验证了该校正算法的有效性，平面两点距离相对误差为0.81%，平面两点方位角相对误差为0.72%，相较于直接利用POS校正误差有显著改善。该研究方法可推广应用于其它面阵摆扫无地面控制点的航飞图像几何校正。

Aiming at the problem of no control point geometric correction for plane array swabbing data, a step by step geometric correction method for plane array swabbing thermal infrared image based on POS data is proposed.The main steps include image correction using the visual vector method, projection to Gaussian auxiliary plane for splicing and virtual control points are extracted. The second geometric precision correction of the global image is carried out by using the least square method to realize the geometric correction of the thermal infrared image in the research area. The validity of the proposed algorithm is verified by obtaining experimental images, the relative error of the distance between two points in the plane is 0.81%, then the relative error of the azimuth angle of two points in the plane is 0.72%, which is significantly improved in comparison with the direct application of POS correction error. The method can be applied to the geometric correction of flight images without ground control points in other plane array swallows.