【作者】 雷盼； 李维嘉； 刘法佑； 吴耀中；

【Author】 LEI Pan;LI Wei-jia;LIU Fa-you;WU Yao-zhong;School of Naval Architecture and Ocean Engineering,Huazhong University of Science and Technology;High-tech Ship and Deep Sea Development Equipment Collaborative Innovation Center;

【机构】 华中科技大学船舶与海洋工程学院； 高新船舶与深海开发装备协同创新中心；

【摘要】 船舶轴系智能装配需要对接法兰的准确空间位置和姿态信息,这些位姿信息可以通过非接触式三维扫描方法获取。由于船舶舱室环境复杂,船舶轴系对接法兰的三维扫描图像存在图像残缺、背景纷杂等问题,因此提出一种从残缺三维扫描图像中提取船舶轴系法兰位姿的算法。首先通过双边滤波算法剔除点云噪声,然后采用GA-BP算法修补残缺点云边界处的数据信息,最后利用kd树查找邻近点、线性形体拟合和三点求取直径等算法,拟合轴系对接法兰圆环面,求取该对接法兰的圆心坐标和法矢量,实现船舶轴系法兰位姿的无接触高精度测量,为轴系后续的快速、自动装配工作奠定基础。更多还原

【Abstract】 Intelligent assembly of ship shafting requires accurate spatial position and attitude information of the butt flange.These pose information can be obtained by non-contact three-dimensional scanning method.Due to the complex environment of the ship’s cabin,the three-dimensional scanned image of the ship shafting butt flange has problems such as defective image and complex background.Therefore,an algorithm for extracting the ship shafting flange position from the incomplete three-dimensional scanned image has been proposed.Firstly,the point cloud noise was removed by the bilateral filtering algorithm,then the GA-BP algorithm was used to repair the data information at the boundary of the residual fault cloud.Finally,finding the neighboring point by the KD Tree,fitting the linear shape and using the three-point diameter,the butt flange torus was obtained,and the center coordinates and the normal vector of the butt flange were obtained to realize the contactless high-precision measurement of the ship shafting flange position and a foundation for the subsequent rapid and automatic assembly work of the ship shafting was laid.更多还原