Hyperspectral Imaging: A Review of Best Practice, Performance and Pitfalls for in-line and on-line Applications

Author:

Boldrini Barbara1,Kessler Waltraud2,Rebner Karsten13,Kessler Rudolf W.1

Affiliation:

1. Reutlingen Research Institute, Process Analysis and Technology, Alteburgstrasse 150, 72762 Reutlingen, Germany

2. Steinbeis Transfer Center Process Control and Data Analysis, Herderstrasse 47, 72762 Reutlingen, Germany

3. Current address: BASF SE, Ludwigshafen, Germany

Abstract

The objective of this paper is to provide a comprehensive review of best practice in hyperspectral imaging. The paper starts to review the taxonomy of the different spectral imaging techniques together with their advantages and disadvantages. The appropriate selection of cameras and spectrographs and their figures of merit are discussed and a detailed description is given of how to qualify and calibrate a pushbroom imaging system for on-line and in-line control. Special emphasis is given to detection and avoidance of specular reflection which can severely distort quantification of the spectral response. Recommendations for an ideal Lambertian illumination are given and the effects of scatter and absorption are discussed when particulate systems are investigated. Here, first principles are introduced and strategies for how to separate scatter from absorption are developed. A simple method using the Kubelka and Munk approach is examined and separated scatter and pure absorption spectra are shown. The same procedure is applied to show the lateral distribution of the separated scatter and absorption properties of an active pharmaceutical ingredient embedded in an excipient. The terms penetration and information depth are discussed and an example of penetration depth profile over wavelengths is provided. Based on a good quality optical set-up and a validated measurement procedure, a practical procedure is described to analyse the data cube using the chemometrics toolbox for hyperspectral imaging. Finally, a survey on selected applications demonstrates the future potential of hyperspectral imaging.

Publisher

SAGE Publications

Subject

Spectroscopy

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3