基于有理函数的遥感水体污染图像的自适应插值

Adaptive interpolation of remote sensing imagery based on rational function model

针对遥感图像的分辨率和显示质量时常不能同时达到需求的问题,提出一种基于有理函数模型的自适应插值算法(adaptive rational interpolation method,ARIM)。该模型不但具有简洁的表达式而且可以在不改变插值数据的前提下通过参数来调整插值曲面的弯曲度。算法充分考虑到了遥感水体污染图像边缘区域的方向信息,让处理过的遥感图像更加符合人眼视觉系统的特性。通过等值线将图像自适应地分解为平滑区域和非平滑区域,在不同的区域分别用不同的算法进行插值,提升了插值速度。另外,模型参数的最优化也有效提高了插值精度。实验结果表明,该算法不但提升了遥感图像的分辨率,而且在主观(目视解释)和客观(峰值信噪比)质量上具有一定的优势。

As the resolution and the quality of remote sensing images can＇t frequently meet the requirements together,we propose an adaptive interpolation method based on the rational function model（ ARIM）,which has an explicit mathematical representation and the shape of the interpolation surface can be modified by adjusting parameters for the unchanged interpolation data. Our new method has given full consideration to the directional information of image edges and divided an image into smooth areas and non-smooth areas by using contour decomposition filter. We employ different interpolation methods for different areas to reduce computational complexity. Besides,the optimization of parameters also improves the interpolation accuracy. Experimental results show that the proposed method not only improves the image resolution effectively,but also improves the image quality in terms of both objective（ PSNR） and subjective aspect.