基于表面法向差分近似的表面重建新算法

Weighted Algorithm Based on Surface Normal Approximation

提出一种基于表面法向差分的四向加权形状重建算法。该方法利用不同表面法向差分方法的重建核函数所具有的互补特性,通过对不同重建结果的四向线性加权可以克服经典重建矩阵的奇异性,并能增加算法的抗噪能力。

Using the inter-complementary speciality of reconstruction kernel function of different surface normal difference approximation, the proposed method combines four-directional reconstruction results through corresponding weighted factors. By means of linear weighting and imposing appropriate boundary condition, the new approach would overcome the oddness of traditional reconstruction kernel function, and is more robust to noise.