Dynamic range extension of a Shack–Hartmann wavefront sensor based on an image processing and sorting method

The accurately measuring strongly aberrated wavefront with a Shack-Hartmann wavefront sensor is hampered by the restricted dynamic range imposed by sub-aperture boundaries and the sensitivity of conventional centroid algorithms to noise. We report an image processing pipeline that removes these constraints without hardware changes. An image segmentation method based on Laplace operator operations integrates Gaussian filtering with morphological processing to generate a binary mask for each spot, demonstrating robustness to substantial displacement. A coarse centroid triggers an adaptive window in which a weighted center of gravity refines the position to sub-pixel accuracy. Row-column dual-sorting method assigns each centroid to its true sub-aperture by enforcing consistency between successive <i>x</i> and <i>y</i> ordered lists, guaranteeing correct wavefront reconstruction. Numerical simulations for Zernike modes <i>Z</i>₄-<i>Z</i>₁₅ show the dynamic range increases 1.13-3.78 times versus conventional COG. At 12 dB SNR, the residual RMS drops from 2.32 to 0.07 µm and the PV from 9.21 to 0.47 µm. Experimental validation with a 27×27 microlens array and an SLM confirms accurate wavefront recovery even when spots lie fully outside their nominal sub-apertures. The method works for both static and moderately dynamic measurements.