alexivaner/Smart-Door-with-Face-Recognition-and-Google-Assistant

This thesis will be focused on the integration of smart door with face recognition and Google Assistant. The smart door will be unlocked if face already recognized but the smart door will stay locked if the face is not recognized and the system will send an email notification to the house owner. The system will be using Raspberry Pi 3 as main microcontroller and servo as locking door actuator. The program will be developed using node-red, Blynk and MQTT platform which are very powerful for developing Internet of Things devices. All of the programs will be coded using Python language that commonly used in Raspbian OS for Raspberry Pi 3. Face detection method will be using Haar Cascade features and face recognition method will using Local Binary Pattern Histogram. Google Assistant integration will use Dialogflow and firebase as Google Cloud services. Integration of Face Recognition and the smart door is successful. Smart door will be unlocked if faces are recognized with an average trust of more than 60%, If the face is not recognized, an email notification containing a face image also successfully sent to the house owner. The Google Assistant could also handle user request successfully with a success rate of 92.8% from 147 trials.