Robust ptychographic imaging with low-discrepancy sequences

Ptychography is a widely-used imaging technique that enables high-resolution, large field-of-view reconstruction via a scanning process. The design of the scanning pattern is critical, as it directly affects both reconstruction quality and computational efficiency. In this paper, we propose a set of new sampling patterns based on low-discrepancy sequences. This non-periodic configuration promotes a relatively uniform distribution of scanning positions and reduces the number of points located far from measured locations. Unlike the common non-periodic scan patterns such as Fermat spiral pattern, which is deterministic but lacks extensibility and exhibits subtle periodicity, low-discrepancy sequences offer provably uniform coverage, support progressive sampling without a predefined number of points, and produce a flatter power spectral density. Consequently, the proposed patterns exhibit enhanced robustness against common experimental artifacts, such as noise and scanning drift. In simulation experiments, we evaluate the performance of low-discrepancy scanning patterns from multiple perspectives, including spatial distribution, power spectral density, and robustness under drifting conditions. Optical experiments further validate the effectiveness and practical utility of our method.