基于k-d树的ICP算法的管道缺陷最深点自动识别研究

Deepest Point Automatic Identification Research of Oil and Gas Pipeline Defects Based on k-d Tree ICP Algorithm

金属管道外表面存在凹坑、腐蚀缺陷等情况,为了准确判断管道的剩余服役寿命,提出一种基于k-d树ICP算法的油气管道缺陷最深点自动识别方法。扫描获取带有缺陷的管道点云数据,提取缺陷处至少1/3管道环向区域点云数据,建立标准圆柱件模型获取点云数据。利用ICP算法对两组点云数据进行配准,基于k-d树算法关联所有无序点云,从而加速搜索点云邻域,快速精确地识别出缺陷最深点。将该算法在天然气长输管段进行验证,以第三方专业检测机构的检测缺陷最深点数据为基准,通过计算对比发现,自动识别方法的误差率仅为0.54%,较之传统人工测量方法,测量误差率降低了3.22%,有效提高了管道外表面缺陷深度测量的准确度。

There are pits and corrosion defects on the outer surface of metal pipeline.In order to accurately judge the remaining service life of the pipeline,an automatic recognition method of the deepest point of oil and gas pipeline defects based on k-d tree ICP algorithm is proposed.The point cloud data of pipeline with defects are obtained by scanning,and at least 1/3 of the point cloud data of pipeline circumferential area at the defect is extracted,and the standard cylinder model is established to obtain the point cloud data.ICP algorithm is used to register two groups of point cloud data,and k-d tree algorithm is used to associate all unordered point clouds,so as to speed up the search of point cloud neighborhood and quickly and accurately identify the deepest defect point.The algorithm is verified in the long-distance natural gas pipeline.Based on the data of the deepest defect detected by the third-party professional inspection organization,the error rate of the automatic identification method is only 0.54%,which is 3.22%lower than that of the traditional manual measurement method.The accuracy of the depth measurement of the outer surface defect of the pipeline is effectively improved.