Dual threshold adaptive update algorithm for eliminating probe edge errors in reflective ptychography

This paper proposes an improved algorithm that optimizes the update range in each iteration of the sample reconstruction process, suppressing the interference of edge errors in each update and preventing erroneous information from propagating into subsequent iterations. While improving reconstruction results, the method does not introduce additional errors, ensures a stable iteration direction, and enhances convergence speed. The algorithm consists of two key steps: global error feedback adjustment and local feature adaptive adjustment. This effectively suppresses the accumulation of edge noise, reduces boundary artifacts, and enhances the recovery quality in weak-signal regions. The proposed algorithm has been validated through numerical simulations and optical experiments. It successfully eliminates artifact effects caused by edge errors in reflective ptychography, providing, to our knowledge, a novel solution for nondestructive testing and imaging of complex samples. The additional step introduces minimal additional computational cost, and the method is experimentally demonstrated in proof-of-concept benchtop experiments. This is a practical improvement to the ptychography reconstruction tool suites and can be easily extended to other ptychography/Fourier ptychography/coherent diffraction imaging reconstruction methods.