Thanks very much Graball.
A quick and simple analysis using parallel lines from the bike to the lane divier markers (which is a valid approach as I found no measurable rotation of the camera throughout the timeframes of interest) shows the biker crossed 2.4 - 2.5 marker repeats (these repeating every 9 meters;
click for 'traffic signs manual'), in the same 0.32 seconds (16NRs) when the gun was taking the measurement.
(yes the gun actually takes a bit longer, I used the wrong frame for the start of the measurement and couldn’t be bothered to correct it - this error had no impact on my own speed deduction, or the relevance of it).Below is an animated gif showing pan-corrected photos. Click on the image to view all of it.
Attachment:
150.gif [322.07 KiB]
Downloaded 2687 times
Below is a close up of how 2.4 markers was arrived at. The green and red lines are as per the animated gif:
Attachment:
2.4 2x.JPG [34.37 KiB]
Downloaded 2112 times
I reckon the repeat distance is 11 pixels for each 9m marker at the point of interest. The green line is 6 pixels from the top marker, and 5 pixels from the next one down. For me that's 2+ (5/11) = 2.45 markers
59NR12, to 59NR28 = (28-12) * 2 / 100 = 0.32s
This yields: 2.4[2.5] * 9 / (0.32) * (3600/1609.3) = 151 [157] mph.
As a secondary check, for a period of slightly over 2 seconds before the biker's brake lights illuminated (the gun measurement was within this time period), the biker crossed 15.0 markers in (2 seconds + 4NRs = 2.08 seconds).
Below is an animated gif showing pan-corrected photos. Click on the image to view all of it.
Attachment:
145.gif [361.85 KiB]
Downloaded 2675 times
This translates to 15.0 * 9 / 2.08 * (3600/1609.3) = 145mph.
(yes it is a bit lower, but the biker would have already eased off in anticipation of the braking).There is no way in the world that speed was 90mph 
IMO, the gun gave a correct speed reading and there's nothing to suggest there could have been any error.