架空线路绝缘喷涂异物检测识别及机器人运动轨迹规划

异物侵入不仅影响输配电线网绝缘漆喷涂，更会影响电网安全运行。为此，需要开展输电架空线路绝缘喷涂异物检测、识别与清除的智能化方案研究。文中构建了基于分段映射直方图图像增强预处理、基于光流法图像区域特征分离提取、基于改进Camshift算法动态异物跟踪识别的异物检测、识别、跟踪方案；构建了基于关节空间的机械臂D-H运动学模型，采用五项式插值法进行机械臂空间运动轨迹规划；构建了基于双层结构模型预测控制(Model Predictive Control, MPC)的异物高精度控制视觉伺服控制方案，实现异物抓取的高精度控制。试验结果表明异物检测、识别、跟踪算法具有高鲁棒性，能够有效增强图像，提取复杂环境下的特征矩形区域，实现有遮挡情况下的动态异物跟踪；多关节机械臂空间运动轨迹算法具有良好连续平稳性，有效满足机械臂始终指向待抓取异物中心点；基于MPC的视觉伺服控制能够满足机械臂异物抓取的高精度控制。

Detection and identification of foreign matters in overhead line insulation spraying and robot trajectory planning

Foreign matter intrusion will not only affect the spraying of insulation paint for transmission and distribution network, but also affect the safe operation of power grid. Therefore, the intelligent scheme of detection, identification and removal of foreign matters in insulation spraying of transmission overhead lines is studied. Firstly, a foreign object detection, recognition and tracking scheme,which based on segmented mapping histogram image enhancement preprocessing, image region feature separation and extraction based on optical flow method and dynamic foreign object tracking and recognition based on improved CAMSHIFT algorithm,is constructed. Then, the D-H kinematics model of the manipulator based on joint space is constructed, and the pentanomial interpolation method is used to plan the spatial trajectory of the manipulator. Finally, a visual servo control scheme for foreign object high-precision control based on double-layer structure model predictive control (MPC) is constructed to realize the high-precision control of foreign object grasping. The experimental results show that the foreign object detection, recognition and tracking algorithm has high robustness, can effectively enhance the image, extract the characteristic rectangular region in complex environment, and realize dynamic foreign object tracking under occlusion; The spatial trajectory algorithm of multi joint manipulator has good continuity and stability, which can effectively meet the requirement that the manipulator always points to the center of the foreign object to be grasped; Visual servo control based on MPC can meet the high-precision control of foreign object grasping of manipulator.