基于实数编码遗传算法的二次曲面提取

Parameters Extraction of Quadric Surface Based on Real-Coded Genetic Algorithm

在机械产品的逆向工程中,针对测量数据提取常用二次曲面是非常重要的。它不仅有助于理解设计意图、指导数据分割,而且可以提高重建模型的精度和速度。首次提出利用基于实数编码的遗传算法针对离散数据点进行二次曲面提取。通过数值实验结果及与其它算法的结果对比,证明了此算法的有效性、鲁棒性和精确性。为逆向工程中基于测量点的二次曲面提取提供了一种新的方法。

Extracting quadric surface from measured data is an important problem in reverse engineering of mechanical parts. It helps not only to understand the designer's intent and guide data segmentation, but also improves accuracy and efficiency of parts reconstruction. This paper proposes for the first time to use genetic algorithm based on real-number-coding to extract quadric surface. Numerical experiments and comparing with other algorithm show the high efficiency, robustness and accuracy of this algorithm. A new method is offered to extract quadric surface based measured points for reverse engineering.