Radiated emissions from a board the size of a Pi are generally about 10% down to the board, and 90% down to the cables connected to it. The VHF band uses and requires an antenna in the order of a metre or two long, and there's nothing on a Pi big enough to be an efficient radiator.
So, make sure that when you test the Pi in your configuration, you have a combination of PSU, enclosure (if any) and cables which closely mirror those of the customers who are reporting interference problems. Pay particular attention to whether or not any cables are shielded, and if they are, exactly where and how the shields are connected at each end.
Without replicating the cable set-up, you won't get anything which even vaguely approximates the amount of RF interference that your customers are seeing. You will see the frequencies that are being generated, but have no way to tell whether the amount of interference at those frequencies will be (a) insignificant, or (b) overwhelming.
In practical terms, the way to work around the problem might be to provide affected users with some good, accurate information about EMC, and how to modify their equipment to improve performance. A screened cable, properly grounded in the correct way, might work wonders.