The next generation radio telescopes such as the Square Kilometre Array (SKA) will be able to produce images of the redshifted Hi (neutral hydrogen) 21-cm signal from the Epoch of reionization (EoR). In this thesis, we use the simulated tomographic images of the signal and investigate methods to extract and interpret information from those images. We implement the image segmentation technique to identify ionized regions in the images and study the statistical information of the EoR in terms of the probability distribution function (PDF), power spectra, bubble size distribution (BSD) etc. We investigate various BSD techniques such as the mean-free-path (MFP), spherical-average (SPA) and friends-of-friends (FOF) to determine the sizes of the ionized regions. We find that the newly implemented method performs better compared to the previously used segmentation methods.

We find that the recovered sizes of the ionized regions are affected by the resolution of the images. As we degrade the resolution to SKA1-Low, the smaller regions are no longer visible and the sharp structures at the edges of the larger regions disappear. Therefore, the observed BSDs will be a modified form of the intrinsic ones. We investigate different line of sight effects on the estimated BSDs. As the ionized regions evolve along the frequency direction in the tomographic images due to the light-cone (LC) effect, the BSDs calculated from the tomographic images are shifted to larger sizes compared to the BSDs from the comoving simulation volume. We find that the error from the comoving boxes in estimating the BSDs is minor if the frequency width of the dataset is less than 10 MHz. On the other hand, another line of sight effect namely the redshift-space distortions (RSDs), induced by velocity gradients along the line of sight, has less than 10 per cent effect at simulation resolution and negligible effect at SKA1-Low resolution on the BSDs.