A Fast T-spline Fitting Method based on Feature Extraction for Large-size Z-map Data

Data fitting is a fundamental tool for constructing a smooth representation from 3D data (Z-map data) in computer-aided design, computer graphics, and reverse engineering. T-spline has been widely adopted for complex data fitting with the advantages of fewer control points, local refinement, and watertight representation. However, the T-spline fitting for Z-map data is inefficient by using a traditional two-phase iterative method, which requests updating T-mesh and recomputing all control points in each iteration. Hence, the traditional T-spline fitting method is time-consuming for large-size Z-map data reconstruction that is widely used in high resolution image processing, geographic information system, and scientific data visualization. In this paper, a fast T-spline fitting method is proposed based on feature extraction for large-size Z-map data. Feature extraction is introduced to construct the ultimate T-spline control grid based on T-spline local refinement without iteration, and an efficient progressive iterative fitting method is employed for T-spline control points evaluation. Computing costs can be reduced obviously since the proposed method is a single-phase iterative method. The proposed method is demonstrated using two types of large-size Z-map data.