该文分析了亚像元技术及相关的几种传统插值方法。针对传统插值方法拟合精度不高的缺点,提出了对两帧错半个像元的遥感图像进行小波插值与滤波的方法,即利用图像中已知点的全局信息对小波系数进行逐层估计。最后通过信息融合,将图像重构为一幅信息更丰富、分辨率更高的遥感图像。由于考虑到图像的全局信息,该算法对图像细节信息保持较好。实验结果证明了该算法优于传统方法,为采用亚像原成像技术提高航天遥感系统空间分辨率提供了一种图像融合重构的新方法。
The paper introduces the subpixel technology and common interpolations. The subpixel technology is adopted for the reason that the traditional method of interpolation accuracy is not high. Wavelet interpolation is combined with median filter technology. The proposed interpolation is based on two low-resolution images offsetting half a pixel. New wavelet coefficients are estimated by the whole information of original images. The high-resolution remote sensing image is fused and recon- structed by two low-resolutlon images offsetting half a pixel. The wavelet interpolation adopts the whole information of the image. The detail of the image is retained well. The test result proves that the proposed method is better than common interpolations. The wavelet interpolation and median filter technology provides a fusing and reconstruction method for raising the resolution of remote sensing imaging system.