As we know security has become a major concern in modern society. It is achieved by surveillance which is capable of taking brisk actions. This paper presents a new system architecture which provides real-time surveillance where unauthenticated people are differentiated and provides the data flow and storage path. Recognition errors are narrowed as much as possible to make the system work efficiently. The proposed approach exploits the Viola-Jones method for face detection, the LBPH (Local Binary Pattern Histogram) algorithm as feature tracker. This combination of algorithms gives more accuracy and less processing time in the proposed design. User data is stored separately from the trained file to make the system more reliable. The proposed design can be integrated into any place which is concerned about security and prevents the security breaches before they occur.