论文部分内容阅读
基于Ｇｒａｎｕｌｏｍ ｅｔｒｉｅｓ 定义的基础，给出了模式谱的定义．运用基于“开”运算的Ｇｒａｎｕｌｏｍ ｅｔｒｉｅｓ得到纹理的模式谱，并以此作为纹理的特征向量，通过双隐层的人工神经网络分类器进行分类，达到识别的目的．给出了该方法在货币识别上的应用实验结果．
Based on the definition of Granulom etries, the definition of the pattern spectrum is given. Granulom etries based on the “on” operation are used to get the texture spectrum pattern, which is used as the texture feature vector to classify by the double hidden layer artificial neural network classifier to achieve recognition. The application of this method in monetary recognition is given.