提出一种基于径向基函数神经网络 (RBFNN)的工件特征参数提取的方法 ,采用了两个径向基函数神经网络 ,利用第一个RBFNN分别求出工件的边缘点pi 邻域内的顺时针边缘与逆时针边缘与x轴的夹角 ,两边缘夹角小的边缘点pi 被认为是具有高曲率的角顶点 .根据工件边缘曲线的特征 ,建立了各种边缘的曲率符号模型 ,用该模型训练第二个RBFNN ,从而识别具有低曲率的切点和拐点及边缘曲线的类型 .采用神经网络的方法提取工件特征参数 ,能准确地定位特征点 A method based on Radial Basis Function Neural Network (RBFNN) is proposed to extract feature parameters of workpiece. Two radial basis function neural networks are used. The first RBFNN is used to find the clockwise Edge and anti-clockwise edge and the x-axis angle, the two edges of the small angle of the edge point pi is considered to have a high curvature of the vertex.According to the characteristics of the edge curve of the workpiece, the establishment of a variety of edge curvature symbol model, using the The model trained the second RBFNN to identify the points with low curvature and the type of inflection point and edge curve.Using neural network method to extract the feature parameters of the workpiece can accurately locate the feature points