초록 열기/닫기 버튼
Robot visual servo technology enables the visual sensor to complete the external environment interaction and provides an intelligent scheme for robot function implementation, such as autonomous perception and autonomous decision making. Reducing the delay of image processing is key to ensuring the real-time performance of the robot visual servo system. Fog computing can supply closely distributed computing resources for industrial robot. In order to meet the functional requirements of the robot visual servo system, this paper proposes a real-time image adaptive processing architecture based on fog computing. According to the requirements of image processing executing the rational distribution of heterogeneous computing resources, it is necessary to ensure that the visual servo system gets enough computational load and completes fast image processing to increase the iteration speed of robot vision system. In addition, by changing the image resolution and sampling frequency of the video stream on the visual servo platform and performing experimental verification for the proposed real-time adaptive image processing algorithm, the results show that the proposed real-time image processing system based on fog computing can significantly improve the calculation efficiency, control precision, and speed of the robot vision system for different types of images.