Diffraction-field-adaptive computational imaging for static and moving targets

Abstract

Single-pixel computational imaging (SPCI) reconstructs a target image from multi-measurements of the correlation between the target and a series of speckle masks. Accurate acquisition of diffraction fields plays an essential role in image formation of SPCI. Loss and distortions in diffraction fields generally lead to the deterioration of the image quality or the failure of imaging. Here, we propose a diffraction-field-adaptive method that provides high-quality image reconstruction for SPCI. In particular, we introduce an adaptive method to accurately calculate the diffraction field in the reference path, where the calculation width of the diffraction field can change itself automatically to make the sampling satisfy Nyquist sampling theorem. Whether the target moves or keeps static, the diffraction field can be accurately obtained and the distortions during light transmission can be reduced by the proposed method. A SPCI model is established and three numerically experiments are conducted to verify the performance of the model. The results demonstrate that the proposed method can avoid diffraction field loss and distortions in the field calculations, and provides excellent imaging performance for both static and moving targets.