دانلود رایگان مقاله یک بازیابی فاز چند تصویری قوی

abstract

The phase evaluation is not perfect by using three kinds of existing multi-image phase retrieval methods. The amplitude-phase retrieval scheme is employed for initializing the input of the Multi-stage algorithm to obtain a robust result. The synthesized axial multi-image phase retrieval technique is able to be highly accurate convergent with continuous phase distribution of [0, 2?] and random phase distribution of [0, ?], which is also demonstrated to have sharper edge for reconstructed phase map in experiment. This paper will provide a powerful and useful guidance for axial multi-image phase retrieval.

6. Conclusion

To sum up, this paper studies the performance of SBMIR, Multi-stage and APR algorithms under two typical object phase distributions and point out their weakness. We put up a hybrid algorithm to tackle the phase problem. The new hybrid algorithm can not only recover different forms of phase, but also obtains a high-precision amplitude retrieval and significantly reduces sensibility to the initialization. Also, the application range limited by the physical size of object is given for these axial methods. The knowledge is very instructive when applying the iterative schemes. At last, these three algorithms and hybrid approach are performed in experiment, which demonstrates that our hybrid approach enhances the contrast of reconstructed phase. Multiple distance measurements, as a lensless imaging technique, is able to facilitate a compact and simple imaging implementation. With the plug-in of our hybrid approach, the robustness and stability of mutli-distance phase retrieval is actually improved, which could be a useful and effective tool for phase imaging or super resolution techqiue.