Illuminant Influence On The Reconstruction Of NIR Spectra 3371 0deec5231f9b126b3b000000
User Manual: 3371
Open the PDF directly: View PDF 
.
Page Count: 6
| Download | |
| Open PDF In Browser | View PDF |
IS&T's 2003 PICS Conference Illuminant Influence on the Reconstruction of NIR Spectra a a a b Meritxell Vilaseca , Jaume Pujol , Montserrat Arjona , and Francisco Martínez-Verdú a Center for Sensors, Instrumentation and Systems Development (CD6), Dept. of Optics and Optometry, Universitat Politècnica de Catalunya Terrassa, Barcelona, Spain b Dept. Interuniversitari d’Òptica, Universitat d’Alacant Alacant, Spain Abstract can significantly increase their cost. The spectral information included in the NIR region is in general directly related to the constituents of a material. Therefore, it is used as an analytical tool in industry and research, and is known as NIR technology.3 The applications include agriculture, the food industry, medical applications, military applications, the chemical industry etc. Multispectral imaging4-7 allows us to obtain the reflectance or transmittance spectra of samples using conventional CCD camera measurements. This technique uses different acquisition channels from which several images of the analyzed sample are obtained. Because of the different spectral response of the channels, the obtained images hold spectral information of the acquired scene. It is therefore possible to calculate the spectral reflectance or transmittance of the original measured sample. The multispectral imaging based methods need all the spectral variables involved in the acquisition process to be known. These variables are; the spectral radiance of the illuminant used, the spectral transmittance of the filters which define each of the acquisition channels and the spectral sensitivity of the CCD camera. After selecting the CCD camera, we can study which illuminants and filters may be used in order to obtain the best reconstruction results of the set of considered samples. Since the mathematical methods used perform approximations, the factors mentioned may yield different quality reconstructions. In this work we studied the performance of two different spectral reconstruction methods, principal component analysis (PCA)6,8-10 and a non-linear estimation method (NLE),11,12 under several lighting conditions. The considered illuminants were of the blackbody type with color temperatures between 1000 K and 16000 K. By numerical simulation, we analyzed the shape of the optimum filters which must be placed in front of the camera in order to obtain good reconstructions of the reflectance spectra of different samples for the tested illuminants. After that, we studied the influence of the illuminant on the quality of the reconstruction using commercially available filters similar to the optimum filters used in the simulations. In order to recover spectral reflectances or transmittances using a multispectral imaging based technique, it is necessary to know the spectral radiance of the illuminant used to light the samples in the acquisition process. In this study, we analyzed the influence of the spectral distribution of the illuminant on the reconstruction of spectral reflectances in the near infrared region of the spectrum (NIR). We considered a set of 30 textile samples with different spectral reflectance in this region. We tested the performance of a principal component analysis (PCA) based method and a non-linear estimation method (NLE), which allow us to obtain the spectral reflectance of samples in the NIR region from a small number of measurements performed with a CCD camera. Using numerical simulation, we analyzed the number and shape of the optimum filters that need to be used in the acquisition channels in order to obtain good spectral reconstructions under several lighting conditions. Finally, we studied the quality of the reconstructions with a set of commercially available filters which are similar to the optimum filters obtained in the simulations. The results obtained show that the reconstruction does not depend heavily on the illuminant used. This indicates that, with the same set of filters, we can obtain good reconstructions for different types of illuminant. Introduction 1,2 Conventional CCD cameras have maximum spectral sensitivity in the visible region of the spectrum. Nevertheless, CCD cameras with improved response in the near infrared (NIR) are currently manufactured and their spectral sensitivity is clearly significant up to 1000 nm. Therefore, this standard instrumentation can be used in order to obtain spectral information of samples in the NIR region (800 – 1000 nm), which is not usually available with conventional spectrophotometers. Standard spectrophoto-meters normally have their response limited to the visible range and require additional sensors to detect energy coming from the NIR (for example, InGaAs), which 536 IS&T's 2003 PICS Conference Spectral Reconstruction Methods X = Cr ≈ CrM + αCv r1 + βCv r2 + ... + ξCv rq , q99.9 % and RMSE < 1. In a previous work it was demonstrated that these values guarantee acceptable reconstructions of the samples in the NIR region10,15. Therefore, the set of commercial analyzed filters can be used to obtain spectral reflectance curves under the influence of all the analyzed illuminants. Conclusion In this work, we studied the influence of the illuminant on the reconstruction of NIR spectra using multispectral imaging methods. We used principal component analysis (PCA) and a non-linear method (NLE) based on a second order polynomial in order to obtain reflectance spectra in the NIR region using CCD camera measurements under several lighting conditions. The analyzed illuminants were 540 IS&T's 2003 PICS Conference graybody radiators with color temperatures between 1000 K and 16000 K with the same radiant flux for both methods and illuminants with color temperature 3371 K and different radiance values in the case of the NLE method. In the first part of the study, we used five equispaced Gaussian filters in order to reconstruct the spectral reflectance of 30 textile samples. We determined the optimum spectral bandwidth of the filters in order to obtain the best possible reconstruction for each analyzed case, that is, for each illuminant and tested method. According to the results obtained, we analyzed a set of commercially available interference filters (Thermo Corion) and analyzed the quality of reconstruction achieved with these filters under different lighting conditions. The results obtained show that Prec > 99.9 % and RMSE < 1 in all the analyzed cases. This indicates that, with the same set of filters, we can obtain good reconstructions for all the considered illuminants in the NIR region. 6. 7. 8. 9. 10. 11. 12. Acknowledgments This research was supported by the Comisión Interministerial de Ciencia y Tecnología (CICYT) (Spain) under grants TAP-99-0856 and DPI2002-00118. M. Vilaseca would like to thank the Generalitat (Government) of Catalonia for the PhD grant she has received. 13. 14. 15. References 1. 2. 3. 4. 5. G. C. Holst, CCD Arrays, Cameras and Displays, SPIE Press, Bellingham, WA, 1998. G. C. Holst, Solid-state Cameras in Handbook of Optics III, McGraw-Hill, NY, 2001, p. 4.1. J. M. Pope, NIR Gains Continue in On-line Process Applications, Chiltons I&CS, 67, 1994, p. 45. F. König and W. Praefcke, The Practice of Multispectral Image Acquisition, Proc. SPIE 3409, 1998, p. 34. J. Y. Hardeberg, F. Schmitt, H. Brettel, J-P. Crettez and H. Maître, Multispectral Image Acquisition and Simulation of Illuminant Changes in Colour Imaging: Vision and Technology, John Wiley & Sons, Chichester, England, 1999, p. 145. J. Y. Hardeberg, F. Schmitt and H. Brettel, Multispectral Image Capture Using a Tunable Filter, Proc. SPIE 3963, 2000, p. 77. B. Hill, (R)evolution of Color Imaging Systems, Proc. CGIV, 2002, p. 473. J. L. Simonds, Application of Characteristic Vector Analysis to Photographic and Optical Response Data, J. of the Optical Soc. of America, 53, 1963, p. 968. I.T. Jolliffe, Principal Component Analysis, SpringerVerlag, NY, 1986. M. Vilaseca, J. Pujol and M. Arjona, Spectral Reflectance Reconstruction in the NIR Region Using Conventional CCD Camera Measurements, Applied Optics (accepted), December 2002. P. G. Herzog, D. Knipp, H. Stiebig and F. König, Colorimetric Characterization of Novel Multiple-channel Sensors for Imaging and Metrology”, J. of Electronic Imaging 8, 1999, p. 342. G. Hong, M. R. Luo and P.A. Rhodes, A Study of Digital Camera Colorimetric Characterization Based on Polynomial Modeling, Color Research and application 26, 2001, p. 76. A. Albert, Regression and the Moore-Penrose Pseudoinverse, Academic Press, NY, 1972. D. Hanselman and B. Littlefield, Matlab 5. A Comprehensive Tutorial and Reference, Prentice Hall, New Jersey, 1998. M. Vilaseca, J. Pujol and M. Arjona, NIR Spectrophotometric System Based on a Conventional CCD Camera, Proc. Electronic Imaging: Machine Vision Applications in Industrial Inspection XI, 2003. Biography Meritxell Vilaseca completed her BSc Degree in Physics at the Autonomous University of Barcelona in 2000. She completed her Degree in Optics and Optometry at the Technical University of Catalonia in 1996. She is currently enrolled on the PhD program in Optical Engineering at the Technical University of Catalonia. Her work focuses on camera calibration and characterization, industrial colorimetry, color management and imaging. e-mail: mvilasec@oo.upc. 541
Source Exif Data:
File Type : PDF File Type Extension : pdf MIME Type : application/pdf PDF Version : 1.4 Linearized : Yes Encryption : Standard V1.2 (40-bit) User Access : Print, Modify, Copy, Annotate, Fill forms, Extract, Assemble, Print high-res Page Count : 6 Page Mode : UseNone Page Layout : SinglePage Has XFA : No XMP Toolkit : XMP toolkit 2.9.1-13, framework 1.6 About : uuid:cdd2e0a6-6cae-4007-bbe0-8ab53db18f01 Modify Date : 2004:04:12 16:16:26-04:00 Create Date : 2003:04:16 16:35:49-03:00 Metadata Date : 2004:04:12 16:16:26-04:00 Document ID : uuid:8a4e5279-4c2d-42da-968e-abee53d9b2c0 Format : application/pdf Title : Illuminant Influence on the Reconstruction of NIR Spectra Creator : Meritxell Vilaseca*, Jaume Pujol*, Montserrat Arjona*, Francisco Martínez-Verdú** Description : PICS 2003 Author : Meritxell Vilaseca*, Jaume Pujol*, Montserrat Arjona*, Francisco Martínez-Verdú** Subject : PICS 2003EXIF Metadata provided by EXIF.tools