Multispectral Imaging

The multispectral camera

The multispectral camera provides high resolution images. The definition (DPI) depends on the dimensions of the work to be digitized.

Scan Joconde en cours

Mona Lisa scan in progress

List of images produced :

  • Definition : from 300 to 1500 DPI (Pixel per Inch)
  • Image type : TIFF and JPG
  • Resolution : 12000 x 20000 pixels
  • Colors (standards) : CIE LAB D65 and RVB
  • Color image, standard CIE LAB D65
  • Color image, standard RVB
  • False color infrared image
  • Infrared image 800 nm
  • Infrared image 900 nm
  • Infrared image 1000 nm
  • Gray tone image, measurement of light reflection « L »
  • Gray tone image, distribution of color space « a »
  • Gray tone image, distribution of color space « b »


Examples of images produced

Image couleur D65 - Répartition de lumière Fausse couleur

Color image D65, light distribution “L”, color spaces “a” and “b”
False color infrared, Infrared 800nm, 900nm and 1000nm


L.A.M. imaging technique (Layer Amplification Method)

Planche contact Joconde

1650 L.A.M. images of the Mona Lisa

The L.A.M. uses the 13 images of the MultiSpectral Camera.
In addition to high-resolution images, this technique generates 1650 images calculated with mathematical algorithms based on the laws of light / matter interaction.
These images make it possible to highlight underlying information that is invisible to the naked eye and sometimes invisible to Infrared.
It is the essential complement to traditional scientific imaging methods: X-ray, Infrared 1700/2500 nm, UV fluorescence etc.


This technique made it possible to discover new elements about the Mona Lisa :

See L.A.M. imaging videos:


L.A.M. example – 3 different states

LAM Joconde

3 different states inside the pictorial layer of the Mona Lisa in L.A.M. #1346 #1474 # 1298


Virtual paint removal

The spectra measured by the multispectral camera can be reprocessed to eliminate the influence of an old yellowed varnish.
A database of artificially aged varnishes allows an algorithm to simulate and calculate colors without the varnish. The result does not give the original colors but the colors underlying the varnish, which have deteriorated over time.


Mona Lisa paint removal

Devernissage Mona Lisa

Mona Lisa today (left), and after paint removal (right)


Identification of original pigments and colors

The spectra measured by the multispectral camera make it possible to identify certain pigments. It is a complement to traditional studies (XRF, Micro Raman, etc.). Software searches the pigment database for combinations of spectra that approximate the spectra measured on the board.
The identification of pigments is an essential contribution in the authentication protocol of an artwork.

Mire 17ème Pérégo lapis corrigé réduit

Perego mire – 17th century


Mona Lisa original colors

After doing the virtual paint removal and identifying the pigments, it becomes possible to find the original colors.

La Jconde - Couleurs d'origine

Colors of the Mona Lisa today (left), and original colors calculated after virtual paint removal (right)


Publications :

  • P. Cotte, D. Dupraz « Spectral imaging of Leonardo Da Vinci’s Mona Lisa: A true color smile without the influence of aged varnish » in proceedings of IS&T CGIV’06, 311 – University of Leeds, UK, 2006
  • P. Cotte, D. Dupraz « Spectral imaging of Leonardo Da Vinci’s Mona Lisa: An authentic smile at 1523 dpi with additional infrared data », in proceedings of IS&T Archiving’06, Ottawa, USA, 2006.

Abonnez vous à notre Newsletter

Rejoignez notre liste de diffusion pour recevoir les dernières nouvelles.

You have Successfully Subscribed!