We present the angular correlation function measured from photometric samples comprising 1 562 800 luminous red galaxies ( LRGs ) . Three LRG samples were extracted from the Sloan Digital Sky Survey ( SDSS ) imaging data , based on colour-cut selections at redshifts , z \approx 0.35 , 0.55 ~ { } and~ { } 0.7 as calibrated by the spectroscopic surveys , SDSS-LRG , 2dF-SDSS LRG and QSO ( 2SLAQ ) , and the AAOmega LRG survey . The galaxy samples cover \approx 7600 deg ^ { 2 } of sky , probing a total cosmic volume of \approx 5.5 ~ { } h ^ { -3 } Gpc ^ { 3 } . The small and intermediate scale correlation functions generally show significant deviations from a single power-law fit with a well-detected break at \approx 1 ~ { } h ^ { -1 } { Mpc } , consistent with the transition scale between the 1– and 2– halo terms in halo occupation models . For galaxy separations 1 - 20 ~ { } h ^ { -1 } { Mpc } and at fixed luminosity , we see virtually no evolution of the clustering with redshift and the data is consistent with a simple high peaks biasing model where the comoving LRG space density is constant with z . At fixed z , the LRG clustering amplitude increases with luminosity in accordance with the simple high peaks model , with a typical LRG dark matter halo mass 10 ^ { 13 } -10 ^ { 14 } ~ { } h ^ { -1 } M _ { \odot } . For r < 1 ~ { } h ^ { -1 } { Mpc } , the evolution is slightly faster and the clustering decreases towards high redshift consistent with a virialised clustering model . However , assuming the HOD and \Lambda CDM halo merger frameworks , \sim 2 - 3 per cent/Gyr of the LRGs is required to merge in order to explain the small scales clustering evolution , consistent with previous results . At large scales , our result shows good agreement with the SDSS LRG result of Eisenstein et al . ( 2005 ) but we find an apparent excess clustering signal beyond the BAO scale . Angular power spectrum analyses of similar LRG samples also detect a similar apparent large-scale clustering excess but more data is required to check for this feature in independent galaxy datasets . Certainly , if the \Lambda CDM model were correct then we would have to conclude that this excess was caused by systematics at the level of \Delta w \approx 0.001 - 0.0015 in the photometric AAOmega-LRG sample .