介绍一种基于GA-BP学习算法的人工神经网络,利用神经网络具有的自适应性、并行性、鲁棒性以及分类能力强等优势,构造玉米品种学习和识别系统.选用3层BP网络自动识别玉米品种,遗传算法进行粗精度的学习以选取网络权值,用BP算法完成给定精度的学习,克服了传统BP算法收敛速度慢、易陷入局部极小等缺陷.结果表明,提出的GA-BP学习算法有效提高了BP算法的收敛速度.%An artificial neural network (ANN) based on GA-BP learning algorithm was introduced. Using the advantages of neural network such as self-adaptability, parallelism, robustness and classification capability to construct a learning and recognition system of corn varieties. A three-layer BP network was adopted to identify the corn varieties automatically. GA algorithm was used to get the initial connection value of the ANN and then BP algorithm was used to train the neural network definitely in order to overcome the shortcomings of the slow convergence and easy to fall into local minimum. The experimental results showed that the proposed GA-BP algorithm effectively improves the convergence speed of BP algorithm.
展开▼
