中文摘要：

针对遥感中线阵推扫数据采集模式和现有方法重构能力的不足,提出基于分块对角矩阵和TV（Total Variation）算法的二维压缩感知模型和方法。该方法能够使压缩感知的约束函数和目标函数同时包含完整的图像二维信息,是真正完全意义上的二维压缩感知。在不改变传统数据采集模式的基础上,通过分块对角矩阵的后处理,实现二维压缩感知。实验结果表明,该方法使图像的重构效果获得极大改善,SNR提高约8dB。但是该方法不适用于OMP（Orthogonal Matching Pursuit）和BP（Basis Pursuit）算法。该方法促进了定向遥感的发展,如与其他模型和方法结合,能进一步提高重构效果。

英文摘要：

For the data acquisition modes of linear array push-broom in remote sensing and deficiencies of exist-ing methods for data reconstruction, the two-dimensional compressive sensing models and methods based on a block diagonal matrix and TV （Total Variation） algorithm were proposed in this paper. This method enabled the constraint and objective functions of compressive sensing to contain the completed two-dimensional image information And it was totally real two-dimensional compressive sensing. On the basis of the traditional data acqui-sition mode, the two dimensional compressive sensing was realized by the post processing of the block diagonal matrix. Experimental results showed that the image reconstruction effect was improved greatly, and the SNR increased by about 8dB. However, this method was not suitable for OMP （Orthogonal Matching Pursuit） and BP （Basis Pursuit）.