以沿曲线等弧长取n个点的曲率半径组成的n维向量为特征向量来描述曲线 ,并利用神经网络实现了曲线类型的识别。与前人的工作相比 ,本文提出的方法能充分考虑曲线形状对零件冲压成形性的影响 ,符合汽车覆盖件智能CAD中对曲线类型识别的要求 The n-dimensional vector composed of the radius of curvature of n points along the arc of the curve is used as the eigenvector to describe the curve, and the neural network is used to identify the curve type. Compared with the predecessors’ work, the method proposed in this paper can fully consider the impact of the curve shape on the stamping formability of the part, which accords with the requirement of curve type recognition in intelligent CAD of automobile cover