摘要
论文针对工业领域对产品目标三维形态检测的具体应用，提出并研究和开发了平行视双目立体视觉体系中以零交叉检测为主，以特征点检测为辅的图象特征提取方法，实验证明。
This paper presents a methodology to extract features of 3D object by analyzing series images grabbed from binocular vision system for industrial application. The methodology is based on zero-crossing algorithm, with feature point detection algorithm added to improve detecting accuracy and speed on the zero-crossing method. It concludes that image features of 3D object can be extracted reliably.
基金
云南省科委应用基础研究基金
