An Improved Algorithm To Enhance Imaging Quality In High-order Thermal Correlated Imaging

Correlated imaging, which is also called "ghost" imaging, is a novel imaging technique through second-order optical intensity correlation measurement to recover the information of objects. Until now various light sources have been employed in the ghost correlated imaging experiment, including quantum optical, pseudo-thermal, and even true thermal light. Recently studies on the higher-order intensity correlation effects of thermal light show that the visibility can be significantly enhanced as the increase of the order N. The first Nth-order(N ≥ 2) correlated experiment in a lensless scheme was performed by recording only the intensities in two optical paths. However, there was a serious drawback that the quality of the correlated images is much poor of)( Nng for the case of n<N-1 compared to that of n=N-1 for the same sampling number. here)( Nng is the Nth-order correlation function for GI, and n is the number of test(object) beams. Therefore, to obtain the same good image quality, a much longer measurement time(sampling number) is required.An Nth-order correlated imaging experiment with pseudo-thermal light is performed by recording only the intensities in two optical paths in a lensless setup, based on a modified Nth-order correlation function. The experiment focuses on analysing the case of the number of test beams n=1.It is found that this new algorithm can effectively remove the noise background encountered in high-order thermal light correlated imaging, so the quality of the reconstructed images in an Nth-order lensless correlated imaging setup has been greatly enhanced compared to former high-order schemes for the same sampling number. Verified the conclusion that with the same number of samples the quality of the image is worse with higher of the order for the traditional higher-order correlated imaging. In addition, the visibility and signal-to-noise ratio for different high-order images on the sampling number has been measured and compared. It is shown that the SNR increases quickly with increasing sampling number, while the visibility initially decrease rapidly and then gradually level off. It can be also found that, as the order increase, the visibility of the high-order image is increasing while the SNR reduces, when the sampling number is large enough. Finally, it analysed the sum and the gap between the maximum and minmum of correlation function when the order is certain. And found that with a few samples a great background will lead to the decrease of the visibility. When the number of samples is increasing, the fluctuation noise of correlation function will be inhibited, then the visibility tends to be smooth. If the number of the order is increasing, the number of samples needed will increase while the visibility to achieve steady.