I think there a four separate images to save bytes. The scanmodus is set to auto to recognize if the document is color, grayscale, black/white etc. It seems that the scan software separates the different colormodels in different images to save bytes.
This scan was made with a special (and expensive) document scanner. This scanner (in combination with the software) has build in support for detecting barcodes. This is not only a software function because it doesn't work with all (Canon) scanners. It is not possible to separate the bar codes in separate images, there is no option in the software to do this.
We would like to use any scanner who can generate PDF-Files to work with barcodes. I attached to more files. One scanned with a sharp multifunction printer (AR-M276) and one sending as a fax but I think the quality of the fax would be to bad to recognize barcodes.
I know it is the "hard work" to find where the barcodes are in the image but this is what we (and mostly other) needs.
Could you post a little more information on how you could see ICEpdf helping you with Barcode recognition?