Image formation challenges in the MOSAIC platform M.E. Gehm a,b , D.R. Golish a , E.M. Vera a , J. Hahn c , D.L. Marks c , and D.J. Brady c a Department of Electrical and Computer Engineering, University of Arizona, Tucson, AZ; b College of Optical Sciences, University of Arizona, Tucson, AZ; c Department of Electrical and Computer Engineering, Duke University, Durham, NC ABSTRACT The DARPA MOSAIC program applies multiscale optical design (shared objective lens and parallel array of mi- crocameras) to the acquisition of high pixel count images. Interestingly, these images present as many challenges as opportunities. The imagery is acquired over many slightly overlapping fields with diverse focal, exposure and temporal parameters. Estimation of a consensus image, display of imagery at human-comprehensible resolutions, automated anomaly detection to guide viewer attention, and power management in a distributed electronic en- vironment are just a few of the novel challenges that arise. This talk describes some of these challenges and presents progress to date. Keywords: multiscale imaging, image stitching, image formation 1. INTRODUCTION High-resolution, wide Field-of-View (wFoV) imagery has great utility in a variety of defense, security, and medical applications. Acquisition of such images, however, has generally proved problematic. The limit on the resolution arises in a fundamental sense from the size of the coherent aperture (the diffraction limit) and in a practical sense from how close to the diffraction limit the system can be designed and fabricated. It is relatively simple to design and manufacture a diffraction limited system for a small aperture and small field; however, as the aperture and field size grow, so does the required complexity of the optical system. As described in a recent publication, 1 the solution to this challenge is break the longstanding tradition of single-axis optical design. Just as parallel architectures enable digital processor scaling, breaking lens design into hierarchical scales enables camera capacity to increase linearly in objective aperture size. This new strategy, which is discussed briefly below and in much greater detail in the aforementioned publi- cation, 1 provides a clear solution to the optical design and fabrication challenges. It does this recognizing that modern imaging systems are hybrid optical/electronic/processing systems and that each component technology has its individual strengths and weaknesses. By considering a novel architecture, the high-resolution, wFoV challenge is moved from the optical domain into the electronic and processing domains where, we argue, the scaling is more favorable. Nonetheless, the electronic and processing solutions (as well as the optical design strategies that will comple- ment the new techniques) are not yet fully identified and understood. The new architecture introduces a number of new challenges in this domain that we are now beginning to address. This manuscript identifies and discusses a number of the key challenges and discusses our recent progress in the development of a multiscale optical system from within the context of the DARPA MOSAIC project—an effort where multiscale optical design will be used to create a 50 gigapixel imager with a 120 ◦ FoV. Corresponding address: gehm@ece.arizona.edu Invited Paper Visual Information Processing XX, edited by Zia-ur Rahman, Stephen E. Reichenbach, Mark A. Neifeld, Proc. of SPIE Vol. 8056, 80560T · © 2011 SPIE · CCC code: 0277-786X/11/$18 · doi: 10.1117/12.883695 Proc. of SPIE Vol. 8056 80560T-1 Downloaded from SPIE Digital Library on 04 May 2012 to 150.135.114.164. Terms of Use: http://spiedl.org/terms