3D computational ghost imaging via binocular vision

Abstract

Ghost imaging, a new type of active imaging, has attracted much attention because of its unique imaging mechanism. As an important application of ghost imaging, three-dimensional correlative imaging has been developed in recent years. At present, the main method of 3D ghost imaging is to use the time-of-flight method to obtain the distance from the target object to the imaging system. However, this method is not easy to achieve millimeter-accurate distance measurement in the distance range within 1m because of the limitation of the measuring instruments. In this paper, by using binocular stereo vision which is a more intuitive 3D imaging method combined with computational ghost imaging, millimeteraccurate distance measurement can be performed on objects within 50 cm. Binocular 3D ghost imaging performs well at close range and is relatively simple to implement, which may help robots avoid obstacles and medical imaging.