Image following using the feature-based optical flow approach

摘要

Visual servo can be achieved by extracting motion information from the optical flow, defined as the apparent motion of brightness patterns of an image. Since the optical-flow-based approach does not require knowledge of targets in advance, it is suitable for versatile tracking and positioning purposes. For most robotic tasks, steady position of the end-effector with respect to its working environment is essential to successful and efficient operations. In this paper, application of the optical flow technique to following of a moving image is presented. In addition, more accurate values of the optical flow are usually observed along image features. In order to achieve better estimation of the optical flow, feature-based optical flow computation approaches are also proposed. Image following tasks are demonstrated by 3D tracking experiments driven by stepping motors.