Design of Gesture Recognition System Based on OpenCV

As a new type of human-computer interaction, gesture recognition has gradually become an important research direction in the field of smart city construction. The traditional gesture recognition system relies on various types of sensors, which has high requirements for hardware and application environment. The recognition accuracy will be affected by the accuracy and service life of electronic components. Therefore, there is an urgent need for natural and easy-to-understand, people-oriented new recognition technology based on computer platform. The OpenCV used in this paper is an open source computer vision library, which contains a wealth of image processing algorithms, which can be easily and quickly applied to gesture recognition systems. This paper makes a complete exploration of gesture recognition system based on OpenCV. Firstly, this paper studies the advantages and disadvantages of different skin color models for skin color detection, and selects a skin color model that is robust in both normal indoor environment and strong light conditions. Secondly, the gray image segmented by skin color is morphologically processed, and the contour frame of gesture information is taken by Fourier descriptor algorithm to obtain the position and boundary information of hand skin color region. Then, this paper extracts and calculates the characteristics of the collected data, and uses the support vector machine algorithm to analyze the gesture recognition of the image captured by the camera. Finally, the above research modules are effectively collected, and the graphical interface is written by PyQT5 for human-computer interaction. The system design framework and the specific process of implementing the system are given. Then a relatively complete experiment was carried out on the system to verify the accuracy of the recognition.