CAJ | 학술논문

提出一种新的卧式金属罐测量方法，方法基于点云拟合，采用八叉树结构管理海量的扫描数据，将扫描点云按空间位置进行分块，从而加快数据处理的速度，提高效率。提出了将基于RANSAC算法的高斯映射所获取的稳定的圆柱拟合参数作为初始值，代入到圆柱距离函数参数化法再进行非线性最小二乘拟合，保证拟合结果的正确性，与传统的卧式罐检测方法进行比较，证明该方法的可行性。
제출일충신적와식금속관측량방법，방법기우점운의합，채용팔차수결구관리해량적소묘수거，장소묘점운안공간위치진행분괴，종이가쾌수거처리적속도，제고효솔。제출료장기우RANSAC산법적고사영사소획취적은정적원주의합삼수작위초시치，대입도원주거리함수삼수화법재진행비선성최소이승의합，보증의합결과적정학성，여전통적와식관검측방법진행비교，증명해방법적가행성。
This paper introduces an improved method of measuring horizontal metal tanks. This method is based on point cloud fitting, Using octree structure management of huge amounts of datas, Scanning point cloud block according to space position, speed up the speed of data processing, improve efficiency. Gaussian mapping based on RANSAC algorithm is put forward to obtain the stability of the cylindrical fitting parameters as the initial values, plug in to the cylinder distance functions parametric method for nonlinear least squares fitting, ensure the correctness of the fitting results, comparing with traditional horizontal tank testing methods, this method has obvious advantages, this algorithm proves to be feasible.