CAJ | 학술논문

边缘保持是单幅遥感图像的超分辨率重建的关键步骤，边缘检测的精度直接影响着遥感图像边缘保持的效果。针对传统边缘检测算子对噪声敏感，边缘检测效果不理想的缺点，采用经典的分数阶微分G-L定义推导出的差分定义方程，构造了该文的分数阶掩模算子。通过与传统边缘检测算子检测结果的对比试验，结果表明：该算子可以有效地提取遥感图像的边缘信息、对噪声有较好的抑制作用和能获得更高的信噪比。
변연보지시단폭요감도상적초분변솔중건적관건보취，변연검측적정도직접영향착요감도상변연보지적효과。침대전통변연검측산자대조성민감，변연검측효과불이상적결점，채용경전적분수계미분G-L정의추도출적차분정의방정，구조료해문적분수계엄모산자。통과여전통변연검측산자검측결과적대비시험，결과표명：해산자가이유효지제취요감도상적변연신식、대조성유교호적억제작용화능획득경고적신조비。
Edge retention is the key step in super resolution reconstruction of single remote sensing image. The effect of edge retention is directly affected by the accuracy of edge detection. In order to overcome shortcomings that the traditional edge detection operators are sensitive to noise and its effect is not ideal,the fractional definition equation was deduced according to classical fractional order differential G-L definition,and the fractional mask operator was constructed. The experiment results of edge detection show that the fractional differential operator can effectively extract edge information of the remote sensing images, has better anti-noise performance and higher PSNR than traditional operators.