Abstract
In this Letter, we introduce a computational sub-aperture fusion technique for Shack–Hartmann wavefront sensing. By incoherently fusing adjacent sub-apertures to emulate large-aperture imaging, this method significantly enhances the signal-to-noise ratio (SNR) within individual sub-apertures. Crucially, it decouples the target SNR from the sampling window size, thereby overcoming the conventional design trade-off between sub-aperture SNR and spatial sampling rate. Both online and offline adaptive optics (AO) experiments demonstrate that this approach substantially improves sub-aperture SNR for low-contrast solar extended objects as well as faint point sources. Consequently, it enhances the detection sensitivity for low-SNR targets and extends the operational limits of AO systems. As a purely computational approach, it offers dynamically tunable performance with minimal computational cost and latency, enabling straightforward deployment in existing AO systems.