基于改进变尺度混沌优化的自标定位置视觉伺服

An Improved Self-calibration Approach Based on Enhanced Mutative Scale Chaos Optimization Algorithm for Position-based Visual Servo

提出不一种基于改进变尺度混沌优化算法(Enhanced mutative scale chaos optimization algorithm,EMSCOA)的机器人动态自标定位置视觉伺服算法,算法在Mendonca-Cipolla和G.Chesi利用本质矩阵进行自标定的基础上进行了扩展.首先依据3个奇异值的特性在线生成目标函数,在进行动态自标定的同时,完成视觉伺服.算法抛弃了G.Chesi方法中对初值选取极为敏感的梯度下降法,采用改进的变尺度混沌优化算法动态优化摄像机内参数.把混沌变量映射到待寻优的5个内参数区间,通过设置内外两层循环,内循环进行混沌搜索,外循环负责缩小内参数搜索区间,避免了混沌优化在内参数区间的盲目重复搜索,提高了搜索效率.算法同时克服了G.Chesi方法迭代过程中要求选取初值时靠近摄像机内参数真值的限制,并可以通过设置参数范围来精确逼近5个内参数.另外,算法不需要物体精确的三维模型,只需要8个空间固定点坐标信息.仿真结果表明,该方法应用于基于位置的视觉伺服时运算速度快,同时对内参数变化鲁棒性强,实验结果证明了算法的有效性.

An improved self-calibrating method for visual servo based on enhanced mutative scale chaos optimization algorithm （EMSCOA） is proposed in this paper. Our approach introduces an extension of Mendonca-Cipolla and G. Chesi＇s self-calibration for position-based visual servo technique which exploits the singular values property of the essential matrix. Specifically, a suitable dynamic online cost function is generated and minimized using an EMSCOA instead of gradient descent method. First, the chaos variables are mapped to the range of the five intrinsic parameters, and then two cycles are set, in which chaos search is performed in the inner cycle and the range is reduced in the outer cycle, in order to avoid blind and repeated searching of chaos optimization in searching space and improve searching efficiency. Moreover, this method overcomes the limitation that the initial parameters must be selected close to the true value, which is not constant in many cases, and the true value is approached by setting the bounds of the unknown parameters off-line. Besides, this algorithm does not require knowledge of the 3D model of the object. Simulation experiments are carried out and the results demonstrate that the proposed approach provides fast convergence speed and is robust against unpredictable perturbations of camera parameters. Experiment result also has verified the effectiveness of the proposed algorithm.