Applicability of Hadamard relaxation method to MMW and THz Imaging with compressive sensing

Compressive sensing (CS) is widely considered a promising method for millimeter wave (MMW) and terahertz (THz) imaging, especially in security screening. In many real-life application scenarios, a CS reconstruction algorithm has to be simultaneously robust, noise-tolerant and fast to be of practical use. However, a lot of CS reconstruction algorithms are not designed aiming at such overall performance, preventing them to become applicable in commercial imaging systems. Having investigated some CS algorithms, we find that Hadamard relaxation method is a potential candidate for commercial CS imaging. By using MATLAB, we study Hadamard relaxation method focusing on its under-sampling ratio, tolerance to noise and efficiency. Comparisons with several other CS algorithms are made using the available data in references. The results demonstrate that the overall performance of Hadamard relaxation method is among the best for real-life and real-time applications of MMW and THz imaging.