高光谱成像技术在香烟品质检测中的应用

摘要香烟市场中的假烟和劣质烟屡断不绝，由于其生产质量要求及卫生环境达不到国家标准。对消费者者的造成更严重的身心双重伤害。传统的鉴别方法简单易行，但是随着造假技术进步，一般传统的方法很难分辨出质量的优劣。本文利用高光谱成像技术对香烟品质进行初步检测。先将不同品质的烟丝在高光谱成像光谱仪(SOC710-E)下进行扫描成像，在波长(400-1000nm)的光谱范围内，得到多组图谱合一的“三维数据”。由于高光谱图像的每个像素点都存在不同波长下的光谱信息，可以在其图像上取其感兴趣区域(ROI),用软件ENVI4.7进行图像校正，把得到的反射值曲线转换成反射率，初步分析比较各个样本反射率曲线的差异。对得到的反射图谱，利用主成分分析法，确定其具有代表性的特征波段。用单因素分析法(One Way ANOVA)和多重比较法LSD对各个样本特征波段进行差异比较分析，进一步确认差异性。表明高光谱成像技术对香烟品质检测的可行性。试图为烟丝品质检测提供一种新方法。50724

该论文有图11幅，表4个，参考文献18篇。

毕业论文关键词：高光谱成像技术烟丝品质反射光谱光谱分析主成分分析

Preliminary Application of Hyperspectral Imaging Technology for Detecting the Quality of cigarette

Abstract The cigarettes of poor quality are filled in the market ,the hygiene levels cannot reach the national standard .it makes great harm to customers .This paper introduces the general application of hyperspectral imaging technology, the basic using of the corresponding data processing software, and use the technique to make the preliminary evaluation of the quality of cigarette .The instrument of this experiment is hyperspectral technique imager SOC710E (400nm-1000nm).We can achieve the image of 3-D data. The original hyperspectral data processed through the ENVI4.7. We chose cigarettes in different proportion to make five standard samples. After four different times of repeated experiments, five average reflectance curves are obtained. In 400 nm - 700 nm band reflectance curves have certain characteristics and degree of differentiation. Making One Way ANOVA and LSD to test the significance level of these curves. The five curves are significant different through the test. From the reflectance curves, we can achieve the possibility of hyperspectral to test different cigarettes. Furthermore, we provide an effective thinking for inspection. on research with hyperspectral imaging technology.

Key Words: Hyperspectral imaging technology Principal component analysis

The quality of cigarette Reflectivity

摘要--Ⅰ

Abstract-Ⅱ

目录--Ⅲ

图清单---Ⅳ

表清单---Ⅳ

1 绪论1

2 材料与方法---9

2.1实验样本-9

2.2检测设备10

2.3烟丝样本光谱图像采集--11

2.4高光谱图像校正和数据处理12

3 结果与分析--13

3.1不同样本光谱反射率分析比较-13

3.2主成分分析选择特征波段-- 16

3.3方差分析比较--17

4 讨论---19

4.1主成分分析作用--19

4.2实验中的误差和不足19

4.3基于实验结果的推断和预测---19

4.4应用和展望19

5结论19

参考文献19

致谢--22

图清单

图序号图名称页码

图2-1 高光谱成像系统的组成高光谱成像技术在香烟品质检测中的应用:http://www.youerw.com/shengwu/lunwen_54090.html