一种基于特征值分解的测量矩阵优化方法

测量矩阵是压缩感知中一个很重要的部分,为了减小测量矩阵与稀疏变换矩阵的互相干性,从而改善重建质量,本文首先通过测量矩阵和稀疏变换矩阵的乘积构造得到一个Gram矩阵,然后定义了一种基于Gram矩阵非对角线元素的整体互相干系数,推导出整体互相干系数与Gram矩阵特征值之间的关系。在此基础上,我们提出了一个最优化模型,在不改变Gram矩阵特征值和的前提下,让每个大于零的特征值的大小都为它们和的平均值,使得测量矩阵和稀疏变换矩阵的整体互相干系数达到最小,从而优化了测量矩阵的性能。将该方法用在一些已知的测量矩阵上,实验结果中矩阵的优化速度快,并且用优化矩阵所得的图像的PSNR有所提高,表明本文优化测量矩阵的方法在重建效果和优化速度方面都有一定的优势。      

基于非均匀FFT的压缩感知雷达信号快速重构方法

雷达处理是压缩感知理论重要的应用方向之一,基于压缩感知的雷达处理可以降低对回波信号的采样速率要求,并且在部分应用中也可改善处理性能。然而,压缩感知重构算法的计算复杂性限制了压缩感知理论在实际雷达信号处理中的应用,尤其是大尺度雷达数据的处理。本文提出了一种基于压缩感知的雷达信号快速重构方法,利用均匀和非均匀快速傅里叶变换运算实现了常规压缩感知重构算法中的矩阵-向量乘法运算,有效降低了重构算法的计算复杂度,加快了压缩感知雷达信号的重构速度。同时,由于引入了快速傅里叶变换运算,该方法消除了大多数常规重构算法对感知矩阵的存储需求。仿真实验验证了该方法的可行性和高效性。      

融合图像块低维流形特性与解析轮廓波稀疏性的压缩成像算法

基于图像的整体稀疏表示和图像块的局部特性,融合图像块低维流形特性和整幅图像在解析轮廓波表示下的稀疏性两种先验知识,该文提出了一种高质量压缩成像算法。该算法利用迭代硬阈值法和流形投影法重构图像。为减小运算复杂度,该文用多个线性子流形的并集来近似表示包含所有图像块的非线性流形,并根据图像块的主方向进行初始分类后再用稀疏正交变换获得各线性子空间的基。实验结果表明,该文算法的重构图像在峰值信噪比和视觉效果两方面均有显著提高。     

Study on the coalescence mechanism of splitting failure of crack-weakened rock subjected to compressive loads

It is of important significance to study the coalescence mechanism of splitting failure of crack-weakened rock masses under compressive loads. In this paper, a simplified mechanism of crack propagation, in which the crack grows along the direction of maximum principal compressive stress, is proposed. Thus, only mode I is taken into account in the formulation and solution. On the basis of the near crack line analysis method, the elastic–plastic stress field near the crack line is analyzed, and the law that the length of the plastic zone along the crack line is varied with an external loads have been established by the matching condition of the elastic- plastic fields on the boundary, the coalescence stress and the strength properties of rock masses have been determined. The solution is a function of the geometry of the crack array. The results show that the crack coalescence depends on the crack interface friction coefficient, the sliding crack spacing, orientation of the cracks, and the crack half-length. The conclusions are of important significance for rock mass engineering.     

Synthesis and characterization of chitosan-hydroxyapatite composite for bone graft applications

In the current work chitosan-hydroxyapatite composites were synthesized using the method of co-precipitation for bone grafting. The composite samples were characterized using SEM, EDS and UTM. The composites showed a good Ca:P ratio of 2.10, indicating a sufficient level of osteoconductive of the composites. The EDS peaks showed a strong presence of carbon, oxygen, calcium and phosphorous with insignificant impurities. The UTM tests revealed compressive modulus of 80 ?MPa desired for various human bones. Also, grey rational analysis was used to optimize the process parameters. The results thus obtained show an improvement over previously reported results of bone grafts using other materials and procedures. Thus, the work can serve as a viable alternative to traditional bone grafting techniques, such as allografts and autografts, which face problems of biocompatibility and affordability.     

No embedding: A novel image cryptosystem for meaningful encryption

In this paper, we propose a novel image cryptosystem, which enables to encrypt the secret images with a smaller-size cover image. Compared with the existing meaningful encryption methods, our cryptosystem has three advantages: (1) non-embedding encryption, i.e., there isn’t any data embedding into the cover image during the encryption process. (2) Our cryptosystem can simultaneously encrypt multiple secret images with one cover image, which greatly improves the security of secret images. (3) Our cryptosystem can accomplish not only the meaningful encryption, but also the meaningless encryption. Thus, people don’t switch encryption methods when meeting different encryption requirements. Our scheme leverages the popular coupled dictionary learning and compressive sensing techniques to accomplish the whole task. Specifically, we use the coupled dictionaries to build connection between the cover image and the secret image, and apply the compressive sensing to decrypt the secret image. To demonstrate the effectiveness of the proposed cryptosystem, a series of experiments are conducted. Experimental results on gray images and colorful RGB images verify its superiority.     

Systematic Synthesis of Compressive VDTA-Based T-T Filter with Orthogonal Control of fo and Q

Based on sixteen nullor-mirror models of the voltage differencing transconductance amplifier (VDTA) and port admittance matrices of the tow-Thomas (T-T) filter with orthogonal control between the characteristic frequency (f_o) and figure of merit (Q), two different categories of the voltage-mode and transconductance-mode T-T filters are synthesized by the means of the nodal admittance matrix (NAM) expansion method. The category A filter that employs two compressive VDTAs and two grounded capacitors includes four structures, and the category B filter that uses two compressive VDTAs, two grounded capacitors, and one grounded resistor, also includes four structures. These circuits are suitable for integrated circuit manufacture, and their parameters f_o and Q can be orthogonally adjusted with varying the bias currents of VDTAs. After the paper and pencil test is completed, the computer analyses, including alternating current (AC), parameter sweep, Monte Carlo (MC), and noise analyses, are performed to support the synthesis approach.     

Blind single-image super resolution based on compressive sensing

Blind super resolution is an interesting area in image processing that can restore high resolution (HR) image without requiring prior information of the volatile point spread function (PSF). In this paper, a novel framework is proposed for blind single-image super resolution (SISR) problem based on compressive sensing (CS) framework that is one of the first works that considers general PSFs. The fundamental idea in the proposed approach is to use sparsity on a known sparse transform domain as a powerful regularizer in both the image and blur domains. Therefore, a new cost function with respect to the unknown HR image patch and PSF kernel is presented and minimization is performed based on two subproblems that are modeled similar to that of CS. Simulation results demonstrate the effectiveness of the proposed algorithm that is competitive with methods that use multiple LR images to achieve a single HR image.     

Hydroelastic impact of a horizontal floating plate with forward speed

The hydroelastic responses of a horizontal plate impacting with the water at both forward and downward speeds are investigated theoretically. The longitudinal bending behavior of a horizontal elastic plate is approximated by the behavior of longitudinal strips represented as an Euler-beam model. A simplified method of hydroelastic responses of the plate is extended to the cases with forward speed and compressive force, for which the hydrodynamic pressure is found by solving a two-dimensional boundary value problem based on the linearized wave theory. In order to validate the theoretical model, a fully-coupled algorithm in LS-DYNA and the available experimental measurements are used for the predictions of the hydrodynamic pressure and deformations of the horizontal plates impacting with water at vertical velocities. The effects of the forward speed and compressive force which can occur at the bottom of ship ships, are investigated theoretically for the plates with different edge boundary conditions. The critical values of the forward speed and longitudinal compression are discussed regarding the plates with various longitudinal lengths.