Jump to main content
Jump to site search


Scalable calibration transfer without standards via dynamic time warping for near-infrared spectroscopy

Abstract

Calibration transfer is crucial for near-infrared spectroscopy (NIR) to avoid time-consuming and labor-intensive recalibration. Classical methods require standard samples from both the master and slave spectrometers, which is the major bottleneck of large-scale NIR applications. In this study, the calibration cransfer method based on variable penalty dynamic time warping (CT-VPdtw) was proposed, which can handle standard-free calibration transfer well and greatly expand the application ranges of the established model. The corn dataset with standards from Cargill and the standard-free wheat dataset from international diffuse reflectance conference (IDRC) 2016 are used to benchmark the transfer ability. When transferring from M5 to MP5 of the corn dataset, the RMSEP reduced from 0.86 to 0.152 by CT-VPdtw, 0.176 by piecewise direct standardization (PDS) and 0.207 by domain-invariant partial-least-squares (di-PLS) respectively. For the standard-free wheat dataset, CT-VPdtw has significant advantages when comparing with PDS and di-PLS. Moreover, 11 slave spectrometers can be transferred to the master spectrometer with satisfactory RMSEPs easily in the wheat dataset, which demonstrates that CT-VPdtw have potential to perform large-scale calibration transfer. With the good transfer ability, the standard-free and large-scale advantages, CT-VPdtw has the potential to be a widely used method for calibration transfer in near-infrared spectroscopy. It was implemented and is available at https://github.com/HMzhu/CTVPdtw.

Back to tab navigation

Supplementary files

Publication details

The article was received on 29 May 2019, accepted on 12 Aug 2019 and first published on 13 Aug 2019


Article type: Paper
DOI: 10.1039/C9AY01139K
Anal. Methods, 2019, Accepted Manuscript

  •   Request permissions

    Scalable calibration transfer without standards via dynamic time warping for near-infrared spectroscopy

    C. Zou, H. zhu, J. Shen , Y. He , J. Su , X. fan, H. Lu, Z. Zhang and Y. Chen , Anal. Methods, 2019, Accepted Manuscript , DOI: 10.1039/C9AY01139K

Search articles by author

Spotlight

Advertisements