Edge detection based on subpixel-speckle-shifting ghost imaging

In this paper, we propose an edge detection scheme by using a subpixel-speckle-shifting ghost imaging, named SPSGI. In the scheme, a set of subpixel-shifted Walsh-Hadamard speckle pattern pairs with low resolution are used to illuminate an unknown object. The edges information with high-quality and higher resolution of the object could be obtained directly by computing with the detection results from the bucket detector and the subpixel-shifted Walsh-Hadamard speckle patterns. The experimental and numerical simulation results show that the quality of the edge detection results is noticeably improved and the number of measurements is greatly reduced comparing with other edge detection schemes based on ghost imaging. In addition, SPSGI has the advantages of enhancing the resolutions of the edge detection with low resolution speckle patterns, which considerably reduces the requirement of the digital mirror devices' resolutions.