Smart Elevator Cotrol System Based on Human Hand Gesture Recognition

The rapid development of computer vision technology makes human-computer interaction possible, which has a wide range of application prospects. In this paper, we propose a gesture recognition system that can be applied to the operation of smart elevators. It can recognize different gestures of people without touching the buttons and reach the designated floor. The training data set used to train the hand gesture recognition consists of pictures and real-time frames taken by the camera. We utilize gesture segmentation, gesture tracking and other methods to preprocess the image. Then we use CNN to train the preprocessed pictures. At last, We design the user interface for computer and human interaction. The experiment shows 98.1% accuracy of static images.