Research On Mobile Robot Dynamic Target Recognition And Navigation Based On Omni-directional Vision

Comparing with traditional vision technology, the omni-directional vision can obtain all vision information of 3D space larger than hemisphere. So it has greatly developed in the field of computer vision. The detecting and tracking of moving objects are hot issues in the same field. The core mission is identifying and tracking target by analyzing vision data. These two technologies have been applied in intelligent robotics and intelligentized weapons, and the combination of them provides much broader prospects of application.With the support of the 863 National Program of China—Based on the omni-directional mobile robot dynamic visual target detection and navigation(NO. 2007AA04Z229), a car navigation system with omnidirectional vision servo as the core technology has been developed. This system can simultaneously achieve comprehensive functions, such as panoramic observation, landmarks navigation and target tracking functions. The core technology research and prototype design will promote a number of individual technologies and the development of integrated innovation systems. It is very important in both academic and practical aspects.This paper covers the following original points:Research on functional design and realization of omni-directional visual servo components, a new design for omni-directional visual servo has been proposed. It combines comprehensive landmarks navigation with maneuvering target tracking functions. In the process, core parts have been choosen or developed; the corresponding image processing algorithm has also been developed. This design realized functions well with low complexity and less cost.A compound identification and tracking algorithm involving both optical flow and particle filter has been developed. This algorithm can determine the relative position of target to robot under landmark or background coordinates, track maneuvering target under moving background. The algorithm is a highly efficient synthesis algorithm, with strong real-time, and robustness.Developing of omni-directional visual servo with dual-DSP: the servo is a parallel processing system based on dual-DSP and module platform. Using this servo, the nevigator's visual function has been enhanced. It is a new application form that using omni-directional visual in monitoring, navigation and maneuvering target tracking.An embedded mobile robot navigation device based on ARM9 has been developed. The navigation device has a core plate and an extended plate. The algorithm loaded can combine the omni-directional visual data with other sensors', and use the combined data for the navigation decision-making. According to the experments, the device works stable and accurately.