Study On Improving The Efficiency Of Collinear Non-interference Phase Retrieval With Reference Beam Constraint

The transmission of information is inseparable from storage technology.With the development of the internet,the interaction of information is becoming more and more frequent,then the amount of data is growing exponentially.The storage of information faces huge challenges.Now the commonly used storage methods are magnetic storage and optical storage,using hard disks and compact discs for storage.But their data-storage capabilities have grown much less rapidly than data growth.Holographic data storage has a unique three-dimensional volume storage method,which is expected to stand out from many storage technologies with fast data conversion rate and high storage density.The collinear holographic storage system is adopted in this thesis.Compared with the off-axis holographic storage system,it is simpler in structure and less susceptible to environmental vibration.The phase encoding is selected in the encoding method.Compared with the traditional amplitude encoding,the encoding rate and signal-to-noise ratio of the phase encoding are higher.In order to decode the phase accurately and quickly,the collinear iterative Fourier transform algorithm is used to reconstruct the Fourier intensity distribution of the reference beam and the reconstructed beam in a single shot for non-interference phase retrieval.On the basis of using this method,the constraint force is enhanced by increasing the reference beam intensity,correspondingly reducing the amount of reference beam used,and even if only a single pixel reference beam is used,more information can be contained in the same input surface data.It can not only increase the data volume of a single data page,but also improve the efficiency of collinear non-interference phase reconstruction.Only reducing the quantity of the reference beam will lead to an increase in the bit error rate.Therefore,while reducing the quantity of the reference beam,the intensity of the reference beam is increased,which increases the constraint of the reference beam on the recording of the hologram.Experiments with single-pixel reference beam can increase the reference beam intensity within a certain range,which can improve the collinear non-interference phase reconstruction efficiency,and the number of iterations can be controlled within 11 times.When the energy of the reference beam is 4 times that of the reconstructed beam,the reconstruction effect is the best.If the light intensity of the reference beam is too large,the information of the reconstructed beam will be overwhelmed,resulting in more bit errors.In order to reduce the bit error rate,we introduced deep learning for noise reduction,reducing the original bit error rate greater than 5% to less than 1%,and improving the accuracy of phase reconstruction..