近IR测定/神经网络分析在废塑料非破坏辨别中的应用  

Non-destructive Discrimination of Plastic Wastes with Near-infrared Spectra Measurement/Neural Network Analysis

作　　者：孟平蕊[1] 李良波[1] 

机构地区：[1]山东建材学院化工系,济南250022

出　　处：《中国塑料》2000年第5期87-91,共5页China Plastics

摘　　要：介绍采用近IR/神经网络分析迅速辨别多种塑料的方法。在近IR的 1.3～ 2 .3μm范围内 51种塑料约 30 0个试样 ,使用标准化的二次微分波谱数据 ,输入在一个三层视感控器的神经网络元中进行分析 ,结果表明 ,对 50种塑料辨别的平均命中率为77%。The possibility of rapid discrimination of plastics with near-infrared spectra measurements and neural network analysis is introduced and examined.The near- infrared spectra in 1.3～2.3μm wavelength region are measured for about 300 samples of 51 kinds of plastics,and normalized second-derivative spectral data are input in a three-layered perceptron-type neural network.The result shows that the average hitting accuracy is 77% for 51 kinds of plastics.

关 键 词：废塑料 神经网络 近红外光谱 废物处理 

分 类 号：X783[环境科学与工程—环境工程]