The merger of binary neutron stars ( NSs ) is among the most promising gravitational wave ( GW ) sources .
Next-generation GW detectors are expected to detect signals from the NS merger within 200 Mpc .
Detection of electromagnetic wave ( EM ) counterpart is crucial to understand the nature of GW sources .
Among possible EM emission from the NS merger , emission powered by radioactive r-process nuclei is one of the best targets for follow-up observations .
However , prediction so far does not take into account detailed r-process element abundances in the ejecta .
We perform radiative transfer simulations for the NS merger ejecta including all the r-process elements from Ga to U for the first time .
We show that the opacity in the NS merger ejecta is about \kappa = 10 { cm ^ { 2 } g ^ { -1 } } , which is higher than that of Fe-rich Type Ia supernova ejecta by a factor of \sim 100 .
As a result , the emission is fainter and longer than previously expected .
The spectra are almost featureless due to the high expansion velocity and bound-bound transitions of many different r-process elements .
We demonstrate that the emission is brighter for a higher mass ratio of two NSs and a softer equation of states adopted in the merger simulations .
Because of the red color of the emission , follow-up observations in red optical and near-infrared ( NIR ) wavelengths will be the most efficient .
At 200 Mpc , expected brightness of the emission is i = 22-25 AB mag , z = 21-23 AB mag , and 21-24 AB mag in NIR JHK bands .
Thus , observations with wide-field 4m- and 8m-class optical telescopes and wide-field NIR space telescopes are necessary .
We also argue that the emission powered by radioactive energy can be detected in the afterglow of nearby short gamma-ray bursts .