Face recognition is a biometric method that lets a device identify a person automatically, and it can be embedded in a surveillance camera for home security. A conventional camera lets an occupant watch the house remotely but is ineffective alone, since it must be monitored continuously. This study develops a Raspberry Pi system that detects and recognizes faces in real time and e-mails the homeowner when an unknown face appears. Detection uses the Haar Cascade Classifier and recognition the Local Binary Pattern Histogram, chosen because both run on a low-cost board without a deep-learning runtime. The system was built on a Raspberry Pi 3 Model B with a 5 MP camera using Python and OpenCV, enrolling five occupants with twenty grayscale images each. Six conditions were tested: a frontal face, a face rotated 45 degrees, a registered face, a printed photo, an unregistered face under dim and bright light, and two faces in one frame. Each was assessed qualitatively as pass or fail in a single staged session rather than through repeated trials, and the e-mail alert fired whenever an unregistered face was detected. The underlying study reports an aggregate 88.89 percent without stating the number of attempts, correct recognitions, or the formula behind it; this figure is repeated as originally reported and examined as a limitation. Given the small enrollment and the limited, manually staged scenarios, the system should be regarded as a preliminary proof of concept for the proposed detection-and-recognition pipeline rather than a validated real-time home-security solution.