Imaging with Blur: Using Motion to Enable Wide-Field Imaging at the Resolution Limit

Optical imaging often involves suboptimal tradeoffs between image resolution, size, and speed. One can obtain diffraction-limited, high-resolution images at high magnification, but only over relatively small fields or at low frame rate if scanning over a larger field. High-resolution imaging also tends to require complex equipment and calibration procedures that impose further practical limits. 

However, much of the information in any single image (let alone sequence of images) is redundant, and thus, compressive modalities leveraging computational analysis can push these limits. 

Here, we demonstrate a novel technique for reconstructing high-resolution, wide-field images at the diffraction limit through intentional use of motion blur. Whereas motion during image acquisition is often considered detrimental, we show that structured motion enables wide-field imaging at the diffraction limit in cases that are directly relevant to materials characterization and large-scale sensing. Motion during image acquisition embeds sub-pixel spatial information into images, which in turn can be readily recovered using sparse image priors well matched to natural images. We also discuss theoretical limits with experimental designs that eliminate ambiguity to enable reconstruction without priors.