In the $k$-median problem, given a set of locations, the goal is to select a subset of at most $k$ centers so as to minimize the total cost of connecting each location to its nearest center. We study the uniform hard capacitated version of the $k$-median problem, in which each selected center can only serve a limited number of locations. Inspired by the algorithm of Charikar, Guha, Tardos and Shmoys, we give a $(6+10α)$-approximation algorithm for this problem with increasing the capacities by a factor of $2+\frac{2}α, α\geq 4$, which improves the previous best $(32 l^2+28 l+7)$-approximation algorithm proposed by Byrka, Fleszar, Rybicki and Spoerhase violating the capacities by factor $2+\frac{3}{l-1}, l\in \{2,3,4,\dots\}$.