基于畸变校正的高光谱偏振图像配准方法

Hyper-spectral polarized image registration method based on distortion correction

提出基于畸变校正的高光谱偏振图像配准方法,分析高光谱偏振图像成像特点构建畸变校正模型,依据高光谱偏振图像成像规律运算畸变校正模型,获取初始畸变系数,采用优化目标函数计算初始畸变系数得到最优畸变系数,采用该系数畸变校正不同偏振方向的高光谱图像;为提升畸变校正图像质量,采用后向插值算法-立方卷积插值方法灰度重建畸变校正高光谱偏振图像,在相同波段的重建高光谱偏振图像中选取0°、60°和120°偏振方向的图像,将60°偏振方向的图像作为基准图像,剩余方向的图像作为待配准图像,采用SURF算法实现高光谱偏振图像的图像配准。经过实验分析发现,该方法畸变校正后高光谱偏振图像径向畸变误差均方根最小值是0.438652个像素,基准图像和配准图像SSD最低数值分别是151.26和157.62,说明该方法畸变校正效果优、配准精度较高。

A hyper-spectral polarization image registration method based on distortion correction is proposed,and the distortion correction model is constructed by analyzing the imaging characteristics of the hyper-spectral polarization image.The distortion correction model is calculated according to the hyper-spectral polarization image imaging law,the initial distortion coefficient is obtained,and the optimal distortion function is used to calculate the initial distortion coefficient to obtain the optimal distortion coefficient.Using this coefficient to distort the hype-spectral images in different polarization directions.In order to improve the quality of the distortion-corrected image,the backward interpolation algorithm-cubic convolution interpolation method is used to reconstruct the distortion-corrected hyper-spectral polarized image in grayscale.In the reconstructed hyper-spectral polarized image,selecting 0°,60°,and 120°polarized images in the same band and using the 60°polarized image as the reference image,the remaining images as the image to be registered,respectively.Using the SURF algorithm to achieve hyper-spectral polarized image registration.After experimental analysis,it is found that the minimum value of the root mean square error of the radial distortion of the hyperspectral polarized image after distortion correction is 0.438652 pixels,the minimum SSD values of the reference image and the registration image are 151.26 and 157.62,respectively,which indicate that the method has excellent distortion correction effect and high registration accuracy.