第36卷  第7期 西 安 交 通 大 学 学 报 Vol.36 No7
2002年7月

Journal of Xi'an Jiaotong Universtity

Jul. 2002

Visual Tracking with Predictor Based on Hierarchical Neural Network
Xi Wenming,Zhu Jianying
(College of Mechanical and Electrical Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China)
retue.gif (1614 字节)zwb.gif (1647 字节)
Abstract:A new controlling scheme of visual servoing is presented. By this scheme, a Jacobian mapping from task space to vision space is expressed and a visual servoing controlling is formulated. This formula is used to predict the future location of image of object. In this way the region of image processing is diminished and the speed of image processing is improved. In order to avoid the complex calculation of inverse Jacobian mapping, a hierarchical neural network is used to replace this calculation. Simulation results show that the speed of image processing and the precision of tracking is improved by these measured.
Keywords:visual tracking;visual Jacobian mapping;prediction;hierarvhical neural network