Page 130 [130]
OCR
c. Post-processing Reflected Images i. Correction of Visible-reflected Images The workflow for the correction of visible-reflected (VIS) images is designed to carry out correction for the factors affecting visible-reflected images, as summarised in Table 3-1 and discussed in Chapter 1, and using the data collected as a result of the acquisition protocols outlined in Chapter 2. Figure 3-19 shows a schematic of the workflow developed in the nip2 software indicating the transformations/corrections to be applied. bs 00. = 000 Camera output XYZ Flat fielded XYZ Flat fielded sRGB Flat converted from RAW + colour fielded + colour calibrated calibrated Camera Output Icc Icc profile profile Correct for Spectral density Converts VIS image Correct for spatial c fo to linear light (XYZ) inhomogeneities orrect for camera of light source response . (Flat fielding) (colour calibration) Figure 3-19. Workflow for the correction of visible-reflected images in the nip2 post-processing workspace. The workflow is composed of the following steps: 1) Loading the input images converted from RAW: — VIS image (with Macbeth chart) and associated VIS flat-field image (if available) for generating the calibration. — VIS image to be calibrated (can be the same as above image or another image taken with the same camera and under the same lighting conditions) and associated VIS flatfield image (if available). 2) Converting the images to linear light (XYZ) using the embedded ICC profiles. 3) Applying a flat-field correction to the images (if this option is checked). 4) Straightening and extracting the Macbeth chart from the linearized, flat-field corrected (flat-fielded) VIS image used to generate the calibration. 5) Generating a colour calibration matrix: — Sets brightness (luminance) of the image from the Macbeth chart greyscale — Generates a colour calibration matrix from a linear regression. 6) Applying the colour calibration to the VIS image to be calibrated (based on the calibration matrix generated from the Macbeth chart). 7) Applying an output ICC profile which converts the linearized, flat-fielded, colourcalibrated VIS image to sRGB (or RGB16) colour space. 8) Saving the sRGB (or RGB16) flat-fielded, colour-calibrated VIS image. Version No. 1.0 122 Date : 14/10/2013
