自由曲面特征识别的研究

Study on recognizing features from freeform surface

由于利用基于区域分割的方法识别曲面往往会造成曲面模型的过度分割,而利用单纯的基于边缘检测的方法识别曲面有时又不能提供较好的区域结构,基于此给出一种自动识别STL(standard triangulation language)形式的CAD模型自由曲面特征的方法——基于边和区域的混合分割方法。通过检测STL三角网格模型的特征边对模型进行有效的区域分割,再计算这些区域上各点的Gauss(K)和平均曲率(H)来确定区域类型,最终达到识别的目的。经验证,该方法能够发挥上述两种方法的优势,获得更好的分割效果,从而识别复杂的自由曲面特征,因而具有较强的实用性。

Using method based on region to recognize surface sometimes made it partitioned excessly, but utilizing the method of detecting edge did not provide good region structure. So this paper gave a new method of automatic recognizing features from freeform surface of CAD models in STL a hybrid region segmentation algorithm based on edge and region, Through detec~ ting feature edges segmented the STL models effectively, then caculated Gauss （K） and Mean curvature at every point on these region. Much experiments prove that the method has the advantadges of the above-mentioned methods, acquiring better segmentation effection, and recognize the complex features from freeform surface, so it has more practicality,