Compressive ghost imaging in the presence of environmental noise

Abstract

Abstract Compressive ghost imaging (CGI) is a single-pixel imaging technique and has attracted the attention of many researchers in the field. Traditional techniques for CGI are very sensitive to environmental noises. In this paper, we study environmental noises in CGI (including time-varying noise and constant background noise). We propose a modified compressive ghost imaging which combines the normalization and blocking of the DC term of the signal to improve the image quality. Time-varying noises are removed by normalization whereas constant background noise is eliminated by the DC blocking. DC blocking is done by subtracting each data and measurement matrix from the average data and the average of all the measurement matrices, respectively. Also, this subtraction optimizes 0–1 binary patterns for CGI algorithm to satisfy restricted isometry property (RIP). Thus, this denoising method offers a very effective way to promote the implementation of single-pixel imaging in real applications such as imaging low-contrast objects even in the presence of environmental illumination. This method is verified both experimentally and by simulation. The calculated signal to noise ratio is improved significantly in the simulation and experiment.