This is not supported, no. It looks similar to the code used by the US postal service.
You would want to look at implementing Reader in core/. It has a framework for handing you rows of image data and reading off black-white-black pixel counts. A lot of that can be reused. Look at the other 1D implementations to get a sense of it.
The tricky bit is that the format is "slightly" two-dimensional. One row doesn't contain enough info to decode. You would have to find, perhaps, the middle of the barcode using the standard framework and then look above and below it for the rest of the data. That's a little different than how the other 1D scanners work, but not terribly different.
Given how wide the barcode is, this won't work if the barcode is rotated much.
You could look at adapting a 2D reader instead, which are better at finding rectangular regions that may be a barcode. It's a longer road to get that to work; it may or may not be better. It would be harder to write but handle rotation better
Good luck, please share your progress if you pursue it.