LAUSR

Abstract The strong need in life science and medicine research has prompted the development of wide-field, high-resolution and real-time microscopic imaging based on camera array. For large-scale sub-image stitching on the order of billion pixels, it is highly challenging to calibrate the camera array with low-overlapping fields of view (FOVs). To address the challenge, we propose a camera array calibration method to obtain the complete camera model and global mapping coordinates. The method utilizes spatiotemporal structured light encoding provided by a spatial light modulator to establish the correspondence between the physical coordinates of the keypoints in the object plane and their pixel coordinates in the sub-image plane. This achieves fast and accurate calibration of a camera-array-based microscopic system without the requirement of sub-FOV overlap. A series of experiments, including reprojection evaluation and global image stitching on the resolution chart and biological specimen slice, verified the effectiveness of the proposed approach. The method supports calibration of global geometric position with 0.2-pixel accuracy and has a broad potential for application to general camera array-based imaging systems.