The first generation of redshifted 21 cm detection experiments, carried out with arrays like LOFAR, MWA and GMRT, will have a very low signal-to-noise ratio per resolution element (\sim 0.2). In addition, whereas the variance of the cosmological signal decreases on scales larger than the typical size of ionization bubbles, the variance of the formidable galactic foregrounds increases, making it hard to disentangle the two on such large scales. The poor sensitivity on small scales on the one hand, and the foregrounds effect on large scales on the other hand, make direct imaging of the Epoch of Reionization of the Universe very difficult, and detection of the signal therefore is expected to be statistical.Despite these hurdles, in this paper we argue that for many reionization scenarios low resolution images could be obtained from the expected data. This is because at the later stages of the process one still finds very large pockets of neutral regions in the IGM, reflecting the clustering of the large-scale structure, which stays strong up to scales of \sim 120 comoving Mpc/h (\sim 1 degree). The coherence of the emission on those scales allows us to reach sufficient S/N (\sim 3) so as to obtain reionization 21 cm images. Such images will be extremely valuable for answering many cosmological questions but above all they will be a very powerful tool to test our control of the systematics in the data. The existence of this typical scale (\sim 120 comoving Mpc/h) also argues for designing future EoR experiments, e.g., with SKA, with a field of view of at least 4 degree.