一种高空间分辨率遥感影像的快速自适应插值算法

A Fast Adaptive Interpolation Algorithm of High Spatial Resolution Remote Sensing Image

针对高空间分辨率遥感影像图幅较大、地物特征丰富以及边缘信息复杂等特点，提出了一种新的快速自适应插值方法，在较低计算复杂度下，有效改善了高空间分辨率遥感影像的插值效果。新算法按照坐标的奇偶性将待插值点分组，并利用Canny算子获得原始图像边缘；将相邻4个像素所组成的矩形区域按照边缘特点划分为5种类型，并针对各类插值点完成快速插值；利用已插值点和原像素点间的关系对剩余未插值点二次赋值。实验结果表明，新算法不仅具有较低计算复杂度，而且有效改善了传统插值方法产生的锯齿和模糊现象，获得了较高的峰值信噪比，对遥感影像插值算法的研究具有重要实际价值。

For the high spatial resolution remote sensing image, the size is vast, the object features are abundant and the edge information is complex. According to its characteristics, a fast adaptive interpolation algorithm applied to the high spatial resolution remote sensing image is proposed. The new algorithm classifies the interpolative dots according to the parity of the coordinate, and gets edges of the original image by applying the Canny operator. Every rectangular area comprised of four neighboring pixel is classified into five types in accordance with the edge characteristics, and the fast interpolation is finished for different types of the interpolative dots. The relationship among pixel of the original image and the assigned interpolation points are used to assign values to the rest dots. Experimental results show that the new algorithm can not only have a low computational complexity, but also accurately preserve edges and prevent blur. In addition, it can effectively enhance the low peak signal to noise ratio. The new algorithm has high practical value to researches on interpolation algorithm of the remote sensing image.