Development Of A Monocentric Multiscale Imaging System With Sparse Subcamera Array

Wide FOV(field of view)and high resolution imaging plays an important role in the field of surveillance,military,astronomy and so on.Wide FOV and high resolution imaging can capture more objectives and enlarge them to show details.Monocentric multiscale imaging system can realize wide FOV and high resolution imaging with simple design and compact structure.The imaging process is divided into two stages: objective lens imaging and micro subcameras imaging.Finally,a wide FOV and high resolution image can be obtained by stitching the subimages captured by the subcameras.The monocentric design of the imaging system can broaden FOV,and the multiscale structure can correct geometric aberration to improve resolution by computational imaging.Due to the low demand for the objective lens and the subcameras used in the monocentric multiscale imaging system,the complexity and cost are much lower than other optical systems with similar performance.In this paper,we have developed a monocentric multiscale imaging system with sparse subcamera array and studied the image mosaic algorithm based on this system.The main contents of this paper are summarized as follows:1,Scaling laws of traditional single aperture imaging systems are derived and the rule that resolution increases with the size of the lens is revealed.The principles of computational imaging and image deblurring are analyzed.The scaling laws are applied to computational imaging.The factors that limit image resolution of computational imaging are analyzed ending with the conclusion that computational imaging can effectively promote image resolution.2,The design idea and the basic structure of monocentric multiscale imaging system are introduced.The main specifications in monocentric multiscale design are analyzed.The objective lens and subcameras are chosen based on the specifications.The aberrations of the objective lens are analyzed by Zemax simulation.The coupling of parameters between objective lens and subcameras are studied to calculate the distance between them.The secondary imaging system is designed based on the analysis of vignetting and aperture of subcameras,overlapping of sub FOV and arrangement of subcameras.3,The mechanical shell of our principle prototype is designed according to the structure and size of the objective lens and subcameras,optical layout and other design parameters.The machine shell is manufactured by 3D printer and the objective lens and subcameras are assembled to build principle prototype.Images are captured using principle prototype.4,Image mosaic algorithm is studied based on the characteristics of our system.Distortion calibration algorithm is studied based on projection model to correct barrel distortions in the raw images captured by prototype.Checkerboard calibration experiment and resolution test chart experiment of single channel subcamera of principle prototype are carried out and the FOV and resolution of final image is estimated.The edge features of corrected images are extracted using Canny algorithm.The features of subimages in overlapping regions are matched to calculate image registration parameters.The subimages are fused based on registration parameters to obtain the final wide FOV images.Automatic image mosaic program for our imaging system is established using parameters calculated from one single group of subimages.