Blue color intensity (CI) values can be read from a tree ring image created with an inexpensive flatbed scanner. Such values have proved to correlate with both density measurements and temperature data. Though question is if these CI values are always relative to something or if they can be made absolute, i.e. calibrated.

Within the color printing industry, scanners, screens and printing machines have since long been calibrated to achieve consistent and expected results. I.e. so that a photograpic image can be scanned  on any calibrated scanner and then printed on any calibrated printer giving a printed image that looks as near as posssible to the original photographic image.

For calibration of scanners, IT8.7/2 calibration cards are used together with a datafile containing colorimetric data read from that card (or from a copy out of the same batch of cards). With the help of special software like e.g. Lasersoft's "Silverfast" a color calibration procedure can then be performed.

We would expect that if we then scan that same IT8.7/2 calibration card again on that same scanner, then we should get an image that complies with the colorimetric data given for that card. This might be a way to check the quality of the calibration.

Note: If you are using Silverfast, then consider to buy also your calibration cards from silverfast.com The Silverfast calibration cards have a bar code telling what card it is, so that Silverfast can easily look up the right datafile for this batch of cards. When using cards of other manufacture, you have to tell Silverfast what data file to use.

If the scanner had been exactly calibrated all these points should be positioned on a straight line going from (0,0) to (255,255). This is not exactly the case above. Especially the dark blue colors (lower left corner) are too light so they should be made somewhat darker to comply with the data in the colorimetric data file of the IT8.7/2 card. The spreading out of dark blue data points is a bit annoying as we are especially interested in the dark blue values representing latewood blue color intensity.

As shown in the diagram, CooRecorder has calculated a mean value curve for these points based on a polynomial fit function. Such a curve can be used to adjust the blue CI values read from the image. The curve can also be used to adjust the image itself.

For practical blue CI measurements and in this case when already using reasonably good calibration with Silverfast, there is no need for an extra calibration step using CooRecorder.

 
//XYZ —> RGB 
var_X = X / 100        //X from 0 to  95.047      (Observer = 2°, Illuminant = D65) 
var_Y = Y / 100        //Y from 0 to 100.000 
var_Z = Z / 100        //Z from 0 to 108.883 
var_R = var_X *  3.2406 + var_Y * -1.5372 + var_Z * -0.4986 
var_G = var_X * -0.9689 + var_Y *  1.8758 + var_Z *  0.0415 
var_B = var_X *  0.0557 + var_Y * -0.2040 + var_Z *  1.0570 
if ( var_R > 0.0031308 ) var_R = 1.055 * ( var_R ^ ( 1 / 2.4 ) ) - 0.055 
else                     var_R = 12.92 * var_R 
if ( var_G > 0.0031308 ) var_G = 1.055 * ( var_G ^ ( 1 / 2.4 ) ) - 0.055 
else                     var_G = 12.92 * var_G 
if ( var_B > 0.0031308 ) var_B = 1.055 * ( var_B ^ ( 1 / 2.4 ) ) - 0.055 
else                     var_B = 12.92 * var_B 
R = var_R * 255 
G = var_G * 255 
B = var_B * 255 

 
    Public Class CCmatrix 
        Public m(0 To 2, 0 To 2) As Double 
    End Class 
    Public D65TosRGB, D50TosRGB As CCmatrix 
    Public Sub setupConversionMatrixes() 
        D65TosRGB = New CCmatrix : D50TosRGB = New CCmatrix 
        With D65TosRGB   'see http://www.w3.org/Graphics/Color/srgb (see srgb.pdf file in ./dok directory 
            .m(0, 0) = 3.2406 : .m(0, 1) = -1.5372 : .m(0, 2) = -0.4986 
            .m(1, 0) = -0.9689 : .m(1, 1) = 1.8758 : .m(1, 2) = 0.0415 
            .m(2, 0) = 0.0557 : .m(2, 1) = -0.204 : .m(2, 2) = 1.057 
        End With 
        With D50TosRGB  'http://www.brucelindbloom.com/index.html?Eqn_RGB_XYZ_Matrix.html 
            .m(0, 0) = 3.1338561 : .m(0, 1) = -1.6168667 : .m(0, 2) = -0.4906146 
            .m(1, 0) = -0.9787684 : .m(1, 1) = 1.9161415 : .m(1, 2) = 0.033454 
            .m(2, 0) = 0.0719453 : .m(2, 1) = -0.2289914 : .m(2, 2) = 1.4052427 
        End With 
    End Sub 
    Public Function XYZ_To_RGB(ByVal X As Double, ByVal Y As Double, ByVal Z As Double, ByRef R As Double, ByRef G As Double, ByRef B As Double) As Boolean 
        Dim result As Boolean = False 
        On Error GoTo atExit 
        If D50TosRGB Is Nothing Then setupConversionMatrixes() 
        Dim ccM As CCmatrix 
        If p_CI_CalibrationTargetCardIsD50 Then 
            ccM = D50TosRGB 
        Else 
            ccM = D65TosRGB 
        End If 
        Dim var_X As Double = X / 100        'X from 0 to  95.047      (Observer = 2°, Illuminant = D65) 
        Dim var_Y As Double = Y / 100        'Y from 0 to 100.000 
        Dim var_Z As Double = Z / 100        'Z from 0 to 108.883 
        Dim var_R, var_G, var_B As Double 
        With ccM 
            var_R = var_X * .m(0, 0) + var_Y * .m(0, 1) + var_Z * .m(0, 2) 
            var_G = var_X * .m(1, 0) + var_Y * .m(1, 1) + var_Z * .m(1, 2) 
            var_B = var_X * .m(2, 0) + var_Y * .m(2, 1) + var_Z * .m(2, 2) 
        End With 
        If (var_R > 0.0031308) Then var_R = 1.055 * (var_R ^ (1 / 2.4)) - 0.055 Else var_R = 12.92 * var_R 
        If (var_G > 0.0031308) Then var_G = 1.055 * (var_G ^ (1 / 2.4)) - 0.055 Else var_G = 12.92 * var_G 
        If (var_B > 0.0031308) Then var_B = 1.055 * (var_B ^ (1 / 2.4)) - 0.055 Else var_B = 12.92 * var_B 
        R = var_R * 255 : If R < 0 Then R = 0 'seems to be missing in the specification at www.easyrgb.com 
        G = var_G * 255 : If G < 0 Then G = 0 'though apparently their Color calculator has implemented this test on zero. 
        B = var_B * 255 : If B < 0 Then B = 0 
        result = True 
atExit: Return result 
    End Function