Developing Smart Security Camera with LBPH Method and Haar-Cascade Based on Raspberry-Pi and Telegram

In this modern era, technological innovation continues to be developed, allowing new technologies that make human work easier, such as biometric technology. An example of biometric technology that needs to be encouraged in using is facial recognition technology. Reportedly, there were 8,493 home theft criminal cases in 2019. However, many people in Indonesia still use security systems in the form of conventional keys. Based on the problem, a tool was developed that can detect faces to access the door of the house. This face detection uses OpenCv, and image processing uses the LBPH and Haar-cascade methods. The system will recognize the faces that the user has registered. If a face that has not been registered tries to enter the system, the system will take a picture of the face that is trying to enter the system and then send it to the homeowner via a telegram bot. The results of this research show the results of user testing in real time under different lighting conditions. The best condition has an average confidence level of 69.44% with an accuracy value of 96% with an error of 4%.