Direct Georeferencing : Conditions to Achieve Accurate Results

Article2: Video Transcript- Direct Georeferencing : Conditions to Achieve Accurate Results

There are some conditions to proper Direct georeferencing: you simply can’t use any GNSS-Inertial system. It has to fit the application. Let me explain some of these simple conditions:

First of all, the GNSS-INS measurement accuracy needs to fit the sensor’s capability. This means, that if you want a laser point to be 15 cm vertical accuracy on this Earth, you’d better make sure the GNSS-INS can measure a better position than that.

Secondly, the GNSS-INS needs to be calibrated to the sensor. Because the Inertial Measurement Unit physically sits on the sensor, we cannot engineer a perfect alignment. So, calibration allows you to mathematically compensate for this. To be able to calibrate means stability. There’s no movement of the CCD, lenses, mirrors…it’s metric.

Because if your sensor is NOT stable, that means your calibration is irrelevant, and thus, you cannot do direct georeferencing.

Last but not least, the GNSS –INS must be highly time-synchronized with the sensor.
For example, let’s say If I’m logging data, and I want 10 cm horizontal accuracy on this earth. but
if my sensors are 10 milliseconds off-sync and I don’t know that, well that might not sound like a lot, but this is a big problem. If I’m traveling in a plane going 200 km/hr (110 knots), that results in 56 cm of error, especially if I want 10 (cm). Plus, I’m trying to measure 5000th of a degree and my airplane can easily have moved by more than 5000th of a degree, 10 ms ago. I’ve missed the mark completely!

If you’ve followed all three conditions, you’re on a good path for direct georeferencing. But I have some common sense tips.

Remember that airborne GPS results are about as good as 5 cm horizontal accuracy absolute. So if your project asks for 5 cm, you’d better be really careful to mitigate all error sources.

Maybe you should consider using direct georeferencing to initialize your aerial triangulation adjustment.

This leads me to say that yes, direct georeferencing can be used to enhance aerial triangulation adjustments by simplifying flight patterns and reducing the number of ground points. It
is especially useful if you’re unsure if you’ve followed all the right procedures for direct georeferencing.

I find that LIDAR/laser scanning users need to be more careful than aerial camera users, and that may be because LIDAR users don’t have the luxury of aerial triangulation.

Finally, even though you can eliminate all ground reference points with direct georeferencing, I would recommend keeping about 10% of what you used to capture as checkpoints.

This is GeoErnest.