RFID is an automatic identification technology that enables tracking of people and objects. This research focuses on improving context-awareness by associating locations to identities. We propose L-VIRT, an extension of our 3-D range- free localization method for RFID tags that uses additional mobile readers to refine its accuracy. Topological constraints are deduced from connectivity information and combined with virtual landmarks to localize tags. Mobile tags are localized with the static reader infrastructure while the localization of static tags is enhanced with mobile readers. Simulation results show a fine-grained accuracy when readers are densely deployed and an impressive error reduction when few mobile readers are involved.