自适应粒子群神经网络识别种蛋成活性  

Research on Automatic Identifying Fertility of Hatching Eggs Using a Self-adapted PSO Neural Network

作　　者：郁志宏[1] 王春光[1] 张晓芳[1] 张莉[1] 

机构地区：[1]内蒙古农业大学机电工程学院,呼和浩特010018

出　　处：《内蒙古大学学报（自然科学版）》2006年第4期464-467,共4页Journal of Inner Mongolia University：Natural Science Edition

基　　金：内蒙古自然科学基金(200408020809)

摘　　要：提出了一种自适应粒子群神经网络自动识别孵化早期种蛋成活性的方法.通过主成分分析提取孵化种蛋颜色特征,减少了神经网络输入节点数.提出的自适应粒子群优化算法,用于优化多层前馈神经网络的拓扑结构,提高了神经网络的学习质量和速度.实验表明该方法识别种蛋成活性切实可行,识别准确性高,算法具有鲁棒性.A self-adapted PSO neural network for automatic identifying fertility of hatching eggs is given. The primary components of feature parameters are extracted and selected with primary component analysis （PCA）. The structure of multi-layer feedback forward neural network is optimized by improved PSO. Learning quality and training speed of the neural network are improved. The result shows that the neural network model for fertility of hatching eggs detection has high accuracy and efficiency and the algorithm is robust.

关 键 词：粒子群算法 神经网络 孵化种蛋 成活性识别 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]