3D image correlator using optically reconstructed integral plane images

Abstract

ABSTRACT In this paper, we propose an optical method for 3-D image correlator using reconstructed integral plane images. In the proposed correlator based on integral imaging, elemental imag es of the reference and signal 3-D objects are recorded by lenslet arrays and then reference and signal integral plane images are optically reconstructed on the output plane by displaying these elemental images into a display panel. Th rough cross-correlations between the reconstructed reference and the single plane images, 3-D object recognition is performed. The proposed method can provide all-optical structure for real-time 3-D object recognition system. To show the usefulness of the proposed method, optical experiments are carried out and the results are presented as well. Keywords: 3-D imaging, integral imaging, elemental images, object recognition 1. INTRODUCTION Integral imaging (II), which was first proposed by Lippmann in 1908, is one of the most attractive display techniques for 3-D imaging and display [1-6]. It is a method that can collect many different perspectives of the 3-D object using a lenslet array. In general, II system largely consists of two processes: pickup and display. In the pickup process, demagnified images with different persp ectives are recorded by an image sensor such as charge-coupled device (CCD) through a lenslet array. The recorded im ages are called elemental images. On the other hand, display process is reverse of the pickup process. The recorded elemental images are displayed by the use of a lenslet array and a display panel such as a liquid crystal display (LCD). Recently, there have been growing intere sts in 3-D object detection and recogn ition based on the II technique [7-15].In 2001, Matoba et al. studied a real-time optical 3-D object corr elation with multiple perspectives imaging using the lenslet array. And, in 2002, Frauel and Javidi proposed a scheme to estimate the longitudinal depth of multiple objects from elemental images and compute 3-D correlations of objects using 3-D shift information. In 2005, J. H. Park et al. also suggested a 3-D optical correlator using a digitally reconstructed sub-image array. However these methods were based on the correlation characteristics betwee n small low-resolution images (elemental image or sub-image). To improve the correlation performance, a 3- D image correlator using computational II reconstruction (CIIR) technique was proposed to detect the 3-D location coordinates of signal objects in space [10]. In this CIIR-based 3-D image correlator, elemental images of the reference and signal objects were imaged through lenslet arrays and recorded by CCD cameras, from which a set of plane images of the reference and signal ob jects were reconstructed on each output plane using a CIIR technique. The principle of the CIIR process is that 3-D images can be digitally reconstructed on the required output planes by superposition of all of the inversely mapped elemental images by using a pinhole array model [11-15]. Therefore the resultant images for image correlation have high-resolution due to the superposition of entire elemental images. From their correlation, it was able to detect improved location data of signal objects. However, the CIIR technique spends larger computational time and memory as the size of reconstructed plane image increases [12]. Therefore it may be difficult to apply to real-time applications. For practical real-time application, all-optical structure may be required. In this paper, we propose a 3-D image correlator using optically reconstructed integral plane images. In the proposed method, optically reconstructed integral plane images are obtained by performing optical pickup and display process for