Single-pixel edge imaging with gradient Radon spectrum

Abstract

Edge detection has been a fundamental component in many image segmentation and object detection systems. Recently, many researchers have implemented direct edge detection with the single-pixel imaging system. Compared with traditional edge detection methods, it does not need the procedures of acquiring the object image and then performing edge detection on the image. However, many previous works need to shift the primary illumination patterns at least twice to get the gradient information in multiple directions, which leads to the surge of the measurements. In this paper, we propose a novel single-pixel edge imaging scheme by obtaining the gradient Radon spectrum. Although the proposed scheme also needs to shift the primary illumination patterns once to obtain the gradient information of the object, compared with the Radon single-pixel imaging, no additional measurements are required. In addition, with the gradient Radon spectrum, this scheme can also directly obtain the positions of the straight edges which can be used for image correction, target positioning, and so on. The experimental and numerical results indicate that a high-quality edge image of the object can be restored by the proposed scheme, and the positions of the straight edges can be effectively located with the gradient Radon spectrum.