Cylindrical surface positioning task is defined as moving a camera to a certain pose with respect to cylindrical surface. To achieve this task, a fringe projection-based visual servoing method is proposed. First of all, phase shiftingbased fringe projection profilometry (FPP) is used to obtain 3D point cloud of the cylindrical surface. Then, the cylindrical axis is extracted by solving a least square optimization problem without constraints. Afterwards, a visual feature selection method is proposed to construct four features of six axis parameters and corresponding interaction matrix and control law are derived to move the camera to desired pose. Special circumstances when interaction matrix is highly decoupled is also discussed. Experiments on cylindrical axis fitting and cylindrical surface positioning task are performed. The results show the visual servo control method is feasible.
To solve problems including the surface distortion and profile error cannot be considered in robotic off-line programming, and the object parts are often texture-less or even without contour, we propose a new visual servo control method directly using the phase map as the visual features. The high accuracy of phase map ensures the high accuracy of positioning or target tracking tasks. Every entry of the phase map is used as visual features so no complex image processing and feature extracting and matching are needed. The corresponding interaction matrix is deduced and control law is constructed. The validity of the proposed method is proved by simulation.