论文部分内容阅读
为了实现可靠的植物病虫害诊断,提出把人工神经网络和多光谱成像技术结合的方法,并将该方法用于常见的三种黄瓜病害的识别研究。在此基础上,实验采用窄带多光谱成像技术获取患病黄瓜叶面的14个可见光通道和近红外通道、全色通道的多光谱图像。利用BP网络对病斑样本的光谱信息进行学习分类。和14通道训练结果比较,增加850nm的近红外通道和全色通道,使网络的训练时间缩短、预测能力提高。实验结果表明,这种方法对植物进行快速、准确和非破坏性诊断提供可靠的技术支持。
In order to achieve a reliable diagnosis of plant diseases and insect pests, a method combining artificial neural network and multispectral imaging technology is proposed. The method is applied to the identification of three common cucumber diseases. On this basis, we used the narrow-band multispectral imaging technique to obtain the multi-spectral images of 14 visible light channels and near infrared channels and panchromatic channels of diseased cucumber leaves. Using BP network to learn the spectral information of lesion samples. Compared with the results of 14-channel training, the increase of 850nm near-infrared channel and panchromatic channel shortens the training time of the network and improves the prediction ability. The experimental results show that this method provides reliable technical support for rapid, accurate and non-destructive diagnosis of plants.