There are several theories around about what the optimum flight line distance should be based on the IMU drift.
IMU’s require always movement to find themselves in earth centers gravity context and continuing on a straight flight path causes them to drift (away) and end up in loss accuracy.
I have seen flight lines as long as 100 Nautical Miles and the solution at the end of the line is marginal at best.
It seems that the Inertial Measurement Unit manufactures (IGI, Applanix, Leica and others) have improved their systems over the years but there are still issues when the tracks are too long. In addition, if the system is flown long distances from the airport to the project area and collection begins immediately without turns the IMU doesn’t find itself usually until the 2nd flight line collected.
So called „S – turns“ flown prior to data collection will help, in case the airport is e.g. 30 minutes away. As a rule of thumb, this should be the way for higher accuracy jobs. e.g. for the 15 cm vertical and 30 cm horizontal jobs, or especially when collecting LiDAR data.
Helicopter systems don’t have this problem typically, so in case you are one of these chopper guys, I would be interested in hearing what you have to say about this topic.