Self-referencing vibration compensation method in in-situ confocal microscopy

In-situ microscopy provides critical insights into the dynamic responses of materials and devices under operational conditions. However, image distortions caused by environmental perturbations and sample inherent vibrations significantly obscure these observations. This paper proposes a self-referencing vibration compensation method that models vibrations as a superposition of multiple cosine components. By incorporating physical constraints based on the point cloud consistency at irregularly sampled points, the vibration parameters are accurately estimated, enabling the reconstruction of distortion-free images. Experimental results demonstrate that the proposed method successfully restores structural details at the single-pixel scale, indicating its potential to achieve sub-pixel accuracy in vibration compensation, and exhibits robust performance even when confronted with images of high structural complexity. This work offers a versatile and effective solution for enhancing measurement fidelity in in-situ confocal microscopy.