一种基于压缩感知的二维几何信号压缩方法

A 2-D Geometric Signal Compression Method Based on Compressed Sensing

本文给出的压缩方法属于谱压缩方法.谱压缩方法是一种常用的二维轮廓线模型压缩方法.文章从压缩感知的角度解释了谱压缩方法,并提出了基于压缩感知的二维轮廓线模型压缩方法.首先利用二维轮廓线模型Laplace算子的特征向量构造了一组基.二维轮廓线模型的几何结构在这组基下可以被稀疏表达.利用随机矩阵对二维轮廓线模型的几何结构抽样,完成压缩.恢复过程中,通过最优化1–范数,实现几何信号的恢复.实验结果表明,该方法压缩速度快,比例高,恢复效果好,适合对大型数据以及远距离数据进行压缩.

Spectral compression method is a commonly used compression method in the field of two-dimensional contour model compression. This paper explains the spectral compression method from the perspective of compressed sensing and provides a compression method of two-dimensional contour model based on compressed sensing. Constructing a basis using Laplace operator of the two-dimensional contour model, we get the sparse representation of the 2-D geometric signal based on this basis. We complete compressing the two-dimensional contour model by sampling the two-dimensional contour model geometry information based on a random matrix. In the recovery process, we reconstruct the 2-D geometric signal through optimizing 1-norm of the sparse signal. Experimental results show that the compression ratio of this method is high, the restore effect is good, and it is suitable for large-scale data compression.