Color single-pixel imaging via compressive Hadamard sampling

Abstract

Single-pixel imaging (SPI) utilizes light fluctuations based on projected structured light patterns to reconstruct images.  This technique is based on estimating the image spectrum from signals acquired by a photodetector. The resolution of the resulting image is heavily dependent on the number of patterns projected. Thus, the acquisition of the full spectrum requires a long acquisition time. In this study, we demonstrate the application of compressive sensing (CS) on SPI based on Hadamard basis patterns to reconstruct color images from subsampled random measurements. By analyzing the reconstructed images using the peak signal-to-noise ratio (PSNR) and structural similarity index measure (SSIM), high-quality images can be reconstructed at 30% sampling ratio, while reducing data acquisition time by more than a factor of three.