Contactless Lift System using Vision Counter and Speech Command Recognition

The current pandemic has brought about change to every aspect of our lives in the recent years. Adaptation to the pandemic demands change to ensure less congestion in public places and more hygienic operation of facilities in public. Lifts are no exception to the changed norms because they are found in multi-storey buildings in the hearts of busy cities. With hundreds of passengers using a lift every day, the buttons of the lift will carry a potential risk, as well as the environment of the lift. The proposed project aims to change the method of operation of lifts making it operate without physical key presses of lift buttons. The project proposes a new method of lift system using vision counter system and speech command recognition. The system operates the lift by interpreting speech commands which are to be used to control the lift. The method of speech recognition used in this work, utilizes artificial neural networks. The model used for the prediction is a Recurrent neural network-based model, that is trained on a dataset by Speech Commands v0.02. The people counter in this report performs image processing on camera live feed to detect movement of passengers in and out of the lift. The smart counter system is followed by object detection algorithm using a MobileNetCaffe model for the purpose of confirmation of the number, by identifying the human features in the detected images.