Compressed ghost imaging can effectively enhance the quality of original image from far fewer measurements, but due to the non-negativity of the measurement matrix, the recover quality is thus limited. In this paper, singular value decomposition compressed ghost imaging is proposed; First, the singular value decomposition be used to decompose the measurement matrix, and then the optimized measurement matrix and measurements are used to recover the original image. Numerical experiments verify the superiority of our proposed singular value decomposition compression ghost imaging method.
Distributed compressed sensing theory is applied to many practical problems, ECG signal, color imaging, etc. In order to improve the reconstruction accuracy of multi-dimensional signals, this paper applies singular value decomposition to the multi-measure vector problem in DCS, then distributed compressed sensing reconstruction method based on singular value decomposition is proposed. This method can achieve row orthogonality of the measurement matrix and does not affect the design of the reconstruction matrix. Numerical experiments verify the effectiveness of the proposed method, which can significantly improve the reconstruction quality of the signal and the robustness to noise.