一种基于模型采样的镀膜双镜头多光谱相机成像畸变校正方法

A image distortion correction method based on model sampling for coated dual-lens multispectral camera

针对传统畸变校正方法无法完成镀膜双镜头多光谱相机影像畸变校正的问题,提出了一种基于数字畸变模型采样的单波段影像畸变校正方法。采用多片后方交会完成相机检校得到各类参数,建立起原始影像的数字畸变模型;对原始数字畸变模型采样得到各波段数字畸变模型;利用各波段数字畸变模型完成对应波段影像畸变的校正。实验结果表明,数字畸变模型采样法能够有效解决镀膜双镜头多光谱相机单波段影像畸变的校正问题,校正精度较高;通过进一步实验证明,经畸变校正后的单波段影像配准误差达到亚像素级,数据可用性高。

Due to the difficulty that traditional method can’t correct the distortion of coated dual-lens multispectral camera images,a new image distortion correction method based on digital distortion model sampling is proposed.First,the digital distortion model of the original image is established by using the parameters calibrated by spatial resection with multi-images.Then the original digital distortion model is sampled to obtain the respective digital distortion model of each band image.Finally,the respective model is used to correct the distortion of the corresponding band image.The experimental results indicate that the digital distortion model sampling method can effectively correct the distortion of the coated dual-lens multispectral camera images with high accuracy,and further experiments prove that the error of band image registrating after distortion correction reaches sub-pixel level,and the data availability is high.