Regularization of Multiple Curves Growth Building Contour Points Based on Multi-Curve Parameter Fusion
Xijiang Chen,Bufan Zhao,Xianghong Hua,Tieding Lu,Han Zhang,Xunqiang Gong,Emirhan Ozdemir
DOI: https://doi.org/10.1109/tgrs.2024.3487269
IF: 8.2
2024-01-01
IEEE Transactions on Geoscience and Remote Sensing
Abstract:The regularization of point cloud contour points is crucial for the accurate recognition of object boundary information, to the realization of three-dimensional representation, object recognition and segmentation. To express the outline of buildings of different curve structures, a regularization method of multiple curves growth is proposed in this paper. First, We construct multi-curve fitting parametric equations for lines, 3D conics, and circular curves based on global least squares, geometric features, and 3D-2D transformations. Second, we propose a multi-curve growth algorithm of contour points based on multi-curve parameter fusion, and segment the multi-curve points. Finally, the distance between points and multi-curve is used to determine the curve type, and the regularization based on different types of curve points is realized by combining the interpolation principle. Experimental results show that the proposed method is more robust than the least squares, CD spline curve and improved B-spline curve, and the precision, recall and F1-score of the proposed method are higher than 80%. The contour points of linear and curvilinear of buildings can be effectively regularized by the obtained multi-curve fitting model. The proposed method can provide urban planners with more comprehensive information on urban architectural forms and spatial layouts. The source code will be available at https://github.com/xijiangyue04/Regularization-of-multi-curve-parameter-fusion.