Adaptive lens-free computational coherent imaging using autofocusing quantification with speckle illumination

Multi-distance phase retrieval (MDPR) based lensfree imaging is promising for an aberration free and compact biological imaging system. In the MDPR processing, the measurement uncertainty of the sample-to-sensor distance undermines its imaging quality and imposes a heavy workload to achieve a perfect reconstruction. The optimal distance can be searched by using an image sharpness quantification function with a refocused data set, however, the scanning is sensitive to noise and aliasing artifact for MDPR. In this work, we propose an adaptive imaging scheme with the help of a diffuser inserted in the lensfree system. The optimal sample-to-sensor distance is searched by combing speckle imaging with sharpness quantification function. With this speckle-based auxiliary, intensity patterns under a coherent illumination are directly used to achieve an in-focusing image reconstruction. Experiments are given to demonstrate the stability, imaging resolution and optical sectioning for our scheme. This method provides a simple, stable and robust tool for the auto-focusing imaging.