基于稀疏表示模型重建TanDEM-X DEM数据

Reconstruction of TanDEM-X DEM Data Based on Sparse Representation Model

为了解决传统DEM插值方法会导致地形过于平滑、数据精度不高的问题,提出基于稀疏表示重建DEM数据。以最新公布的TanDEM-X DEM为实验数据,利用双线性插值法、克里金插值法、反距离权重插值法和本文方法对TanDEM-X 90m DEM进行重建生成较高分辨DEM,并从视觉检查、统计值和DEM差值图对不同算法进行分析和评估。实验结果表明,本文方法生成的DEM可以避免传统插值方法造成的平滑地形问题,有效抑制地形噪声,地形细节特征丰富,能够较好地保持真实地形。

In order to solve the problem that the traditional DEM interpolation method can lead to the terrain being too smooth and the DEM accuracy is not high,the method based on the sparse representation is proposed to reconstruct the DEM data.Using the recent published TanDEM-X DEM as experimental data,the TanDEM-X 90 m DEM is reconstructed by bilinear interpolation,Kriging interpolation,inverse distance weight interpolation and the proposed method to generate higher resolution DEM.Different algorithms are analyzed and evaluated by visual inspection,statistical parameters and DEM difference maps.The experimental results show that the DEM generated by the proposed method can avoid the smooth terrain problem caused by the traditional interpolation method,effectively suppress the terrain noise,and have rich terrain features,which can better maintain the real terrain well.