In this paper, we address the problem of secure estimation in networked systems, by focusing on false data injection attacks in large-scale systems, where malicious attackers alter the original transmitted data between subsystems. We propose a technique that ensures asymptotic secure estimation of the original transmitted data under two attack classes, termed stealthy and non-stealthy, while also providing detection and isolation capabilities. We give conditions under which asymptotic recovery of nominal performance is guaranteed, thus providing the large-scale system with resilience. Furthermore, we demonstrate the effectiveness of the proposed technique through a simulation-based case study.