基于Kent混沌测量矩阵的压缩感知图像重构算法

COMPRESSED SENSING IMAGE RECONSTRUCTION ALGORITHM BASED ON KENT CHAOTIC MEASUREMENT MATRIX

图像重构是图像数字化和恢复高质量图像信号的关键技术,使用压缩感知理论进行图像重构的意义在于显著减少采样次数,降低系统资源的消耗。测量矩阵的构造是压缩感知的重要研究内容之一。提出一种基于Kent混沌测量矩阵的压缩感知图像重构算法,将Kent混沌序列作为测量矩阵,采用离散小波变换的稀疏化方法,在小波域对原始图像信号进行测量。最后采用正交匹配追踪方法恢复原始图像。仿真实验中,对比高斯随机测量矩阵和Logistic混沌测量矩阵,对不同的图像进行重构。实验结果证明,基于Kent混沌测量矩阵的重构算法能够恢复原始图像,重构性能优于高斯随机观测矩阵和Logistic混沌测量矩阵,同时克服了随机测量矩阵硬件难以实现的缺陷。

Image reconstruction is the key technique of image digitisation and restoration of high-quality image signal. The significance of using compressed sensing theory to reconstruct the image is to reduce the sampling times and decrease the consumption of system resources. The structure of measurement matrix is one of the important research contents of compressed sensing. This paper presents a compressed sensing image reconstruction algorithm based on Kent chaotic measurement matrix. We use Kent chaotic sequence as the measurement matrix and adopt sparse method for discrete wavelet transform to measure the original image signal in wavelet domain. Finally, we use the orthogonal matching pursuit method to recover the original image. In simulation experiments, Gaussian random measurement matrix and Logistic chaotic measurement matrix are compared in the reconstruction of different images. Experimental results show that the reconstruction algorithm based on Kent chaotic measurement matrix can reconstruct the original image, its reconstruction performance is superior to Gaussian random measurement matrix and Logistic chaotic measurement matrix, and the defects which random measurement matrix hardware cannot realize are overcome.