基于BP神经网络的田间杂草识别技术研究

设计了一种以田间除草指标体系作为神经网络的输入,以田间除草等级作为输出的田间除草综合评价模型;以田间除草指标的各级评价标准作为模型的训练样本和检验样本,设计了一种神经网络算法,利用Matlab软件对BP神经网络进行训练和检验.结果表明:BP神经网络对检验样本的模拟输出和期望输出是一致的;BP神经网络人工智能技术应用到田间除草,具有运算速度快、精度高,过程方便简捷的优点.

An evaluation model for the field weed control based on BP nearal network

This paper proposed a comprehensive evaluation model for the field weed control. In this model, the input and the output of the neural network are the field weed control index sys tem and the weed control level, respectively. We designed a neural network with the training sample and test sample of the model as the evaluation standards of every weed control level, then trained and tested it by Matlah software. The results showed that the consensus can be realized in both the analog output and expected output of the BP neural network. This model has the charac teristics of fast speed, high precision and simple process in the field weed control.