Oldal 139 [139]
OCR
ee @ < CHARISMA I 7 23) The 3x3 calibration matrix M is generated by comparison with the uploaded .mat file (Figure 3-32) and the rows below show the average AE (colour error) and list the worst colour match. Note that AE values should not be greater than 5 to be deemed an accurate calibration. Applying the colour calibration to the VIS image to be calibrated (based on the calibration matrix generated from the Macbeth chart). 24) Scroll right to column H, which applies the colour calibration to the VIS image to be calibrated. This can be the same image as was used to generate the calibration (Figure 3-33) or another image taken with the same camera and under the same lighting conditions (Figure 3-34). The colour calibration matrix generated in column D is applied to the image in the “object” thumbnail from the “linear” tab (J13). 25) The result is shown in the “object” thumbnail at H10 (Figure 3-33 and Figure 3-34). Elle Edit View Toolkits Help | Input linear markup | viscalib | uvicalib align ! specden falsecolour uvistray uvikm vilstray vikm results + V H- calibrate visible reflectance image a H10 label | visible reflectance S| object | Apply. calib 08 linear.313 sample Oo \ macbeth | Pick a Macbeth data file: macbeth_lab_d65.mat A v iS mode | Input LUT: Linear input, set brightness from chart |< | A v M )0.290469. (0.674184 0.0158778 -0.543422 |1.51308 — :0.0303436 "0.180834 0.143996 | 1.04073 scale | 1.325384 avg_dé76 | 4.992412 worst patch | Blue (patch 13, 13.174 de) input Find_calib_item.action D1 ( IGA 57 Ham 6B o] fő 5.| Be) Bel 0.] jo EEl[ seach deste [SOO GS wa Figure 3-33 26) Scroll right to column C, which uses the position of the Spectralon diffuse reflectance standards in the VIS image defined in the “markup” tab to extract their position in the Calibrated VIS image at H10. Note that this step is required for later workflows and can be omitted if only correcting visible-reflected images. A matrix of values is created by subdividing the image of the standards by the number entered and determining the average X, Y and Z values for each of these (Figure 3-35). The sampling visualisation tool and slider are again provided here (not shown). Version No. 1.0 131 Date : 14/10/2013
