In this paper, we investigate the coverage problem from the perspective of target localization for wireless camera sensor networks. We first propose a novel localization-oriented sensing model based on the perspective projection of camera sensors. Based on the sensing model, we propose a new notion of coverage, Localization-oriented coverage (L-coverage for short), by using Bayesian estimation theory. Furthermore, we analyze the relationship between the density of camera sensors and the L-coverage probability under random deployment where camera sensors are deployed according to a 2-dimensional Poisson process. According to the relationship between the density of camera sensors and the L-coverage probability, we derive the density requirements for an expected L-coverage probability. We validate and evaluate our proposed models and schemes by simulations.