应用多结构神经网络建立了大米外观品质评判模型，可实现5类大米的识别。模型采用5个并行工作的多层前向神经网络。每个多层前向神经网络包含两个隐含层，以大米图像的形状特征和颜色特征作为网络输入。网络训练和仿真结果显示模型识别的平均准确率为92.66%，比相同网络复杂度下的多层前向神经网络模型提高5.04个百分点，并且网络学习速率快。
A multistructure neural network (MSNN) was proposed and applied to classify five classes of rice grains. The MSNN model consisted of five parallel multilayer feedforward neural networks (MLNN). With two hidden layers MLNN was trained using morphological and color features of the rice grains extracted from their images as input. The average classification accuracy of MSNN was 92.66%, with an increase of over 504 percent points than that of MLNN; moreover the network training time for MSNN was shorter than that for MLNN.
[1]谢 健. 我国大米标准的现状及修订思路. 粮食与饲料工业, 2006(4): 6-9.
[2]孙 明, 石庆兰, 孙 红, 等. 基于计算机视觉的大米外观品质检测. 沈阳农业大学学报, 2005, 36(6): 659-662.
[3]黄星奕, 吴守一, 方如明, 等. 遗传神经网络在稻米垩白度检测中的应用研究. 农业工程学报, 2003, 19(3): 137-139.
[4]陈建华, 姚 青, 谢绍军, 等. 机器视觉在稻米粒型检测中的应用. 中国水稻科学, 2007, 21(6): 669-672.
[5]Gupta L, Upadhye A M. Nonlinear alignment of neural net outputs for partial shape classification. Pattern Recog, 1991, 24(10): 943-948.
[6]Ghazanfari A, Irudayaraj J, Kusalik A. Grading pistachio nuts using a neural network approach. ASAE, 1996, 39(6): 2319-2324.
[7]de Villiers J, Barnard E. Back propagation neural nets with one and two hidden layers. IEEE Trans Neural Networks, 1992, 4(1): 136-141.
