On July 1, 2020, a paper entitled “Compact three-dimensional computational imaging using a dynamic virtual camera” was published online in Optics Letters, a leading general optical journal in the world. The paper about the latest progress in the three-dimensional (3D) imaging field was authored by Prof. Anhu Li (first author and corresponding author), PhD student Xingsheng Liu and graduate student Zusheng Zhao from School of Mechanical Engineering, Tongji University.

The researchers originally propose a 3D computational imaging technique using a dynamic virtual camera. By controlling the smooth rotation of an optical element for the continuous adjustment on real camera boresight, a dynamic virtual camera is formed to move in a certain path. The virtual camera could observe and record full-field high-resolution object images from an infinite number of viewpoints. Since the motion law and instantaneous location of the virtual camera are available, a stereo matching method based on multiple geometric constraints can be developed for multi-view image registration. On these basis, an object or scenery in space can be recovered through backward ray tracing and 3D computational reconstruction to provide the desired 3D coordinates, spatial profile and texture information (Fig. 1).

Fig. 1. Imaging principle of a dynamic virtual camera.

The experimental results has proven that the proposed technique can provide lager field of view, higher imaging resolution and more sufficient view angle. In addition, the imaging system is advantageous in compact structure and flexible parameter configuration, which achieves 3D reconstruction accuracy competitive to conventional stereovision. It is believed that the dynamic-virtual-camera-based 3D imaging technique may be a good alternative to the existing approaches.

Fig. 2. 3D imaging performance of a dynamic virtual camera.

Prof. Anhu Li’s research group is engaged in photoelectric imaging tracking technology and its measurement applications. Recently they have also made significant progress in visual tracking and precision measurement, which has been published in the mainstream journals in this field. Their latest achievements include cooperative camera multi-scale surveillance method (Precision Engineering, 2020, in press), inverse solutions to submicroradian laser tracking problem (Journal of Lightwave Technology, 2020, 38(6)), and variable-boresight visual tracking model for industrial robots (Optics Express, 2020, 28(4)). The above work was funded by National Natural Science Foundation of China (61975152, 61675155, 51375347).

See more details in https://www.osapublishing.org/ol/abstract.cfm?uri=ol-45-13-3801.