Embedded Files
For each spectral band i, henceforth called “wavelength”, a standard acquisition process is followed. A white, highly reflective target is placed in front of the object. “White” image, W, is acquired. The white target is removed and, keeping all settings unchanged, the acquisition of spectral image Q follows. Still keeping settings unchanged, light entrance is blocked in front of the camera and a “Black” image, B, is captured, corresponding to the dark current of the camera.
Intensity normalization is performed independently for each spectral image and for each pixel x as:
The procedure of image manipulation is done with the free, open source, software ImageJ.
ImageJ
Procedure
Step 1
For each wavelength (filter) drag and drop the three image files at ImageJ (white, picture, black).
Now you will have something like that.
Step 2
In order to remove the noise (black image) from our painting go: Process->Image Calculator. A small window will pop up.
Step 3
Image1: The image of the painting.
Image2: The black image.
Operation: Substruct.
Click Ok.
The result of this manipulation will be shown in a new window with a new image name.
Step 4
In order to remove the noise from the white image select again Process->Image Calculator. And in the small window select as image1 the white image.
The result of this manipulation will be shown in a new window with a new image name.
Step 5
In order to get the normalised image select again Process->Image Calculator. And in the small window select as image1 the Result of the manipulation of the painting and as Image2 the result of the manipulation of the white image. As operation choose Divide and select the option "32-bit (float) result".
Step 6
The normalized image will pop up. You can save this image if you select: File->Save as->BMP. A good practice is to save all normalized images in a specific folder. You can do these steps for each wavelength (filter) in order to get all normalized images.
Google Sites
Report abuse