AERIAL AND LIDAR SURVEYS – GROUND CONTROL PITFALLS
General
In recent years advances in Aerial Technology made Aerial and Lidar surveys much more affordable. Combining IMU`s (Inertial Measurement Units) and GPS units with existing Photogrammetric technology reduced the number of Ground Control points required on a typical Aerial Photography or Lidar Mapping survey by as much as 90%. However, it is important to remember that the basis of Aerial or Lidar mapping is still the ground control and as a result, accuracies are limited to the accuracies of the ground control survey.
Accuracies
There are a number of ways to describe accuracy and precision of an Aerial/Lidar survey, and care should be taken to understand the differences between the various methods.
Per definition
| Precision | The ability of a measurement to be consistently reproduced. |
| Accuracy | The degree of closeness of measurements to its actual (true) value. |
| RMS | This is the standard deviation of a list of measurements, a measure of the deviation of values from the mean of the measurements. Please note that it is a measure of the deviation from the mean of the measurements, and not a measure of the deviation from the true value. |
Accuracy and precision can be illustrated by measuring the length of a 1m steel beam using a measuring tape with 10cm`s cut off the end. If the measurement is done 10 times there will be little difference between the 10 measurents, i.e. the measurements will have high precision. However, all the measurements will be incorrect by 10 cm`s and the accuracy is therefore poor.
How can the accuracy be determined? The accuracy can be determined by comparing results from the Aerial/Lidar survey with known and true positions on the ground. Known and true positions can be determined using a number of different conventional survey techniques, most often GPS (for horizontal) and Spirit Levelling (for vertical).
Heights
It is very important to distinguish between Ellipsoidal and MSL (Mean Sea Level) heights and the derivation of MSL from Ellipsoidal heights. Ground control for Aerial/Lidar surveys is often done using GPS equipment. GPS devices take measurements relative to the WGS-84 reference coordinate system – a mathematical ellipsoid of the earth. Mean Sea Level is a gravitational equipotential surface (an uneven surface over the earth with equal gravity).
On most Aerial and Lidar surveys the required height datum reference is MSL, and as a result, Aerial and Lidar companies have to convert or relate their data to MSL. How do they achieve this? The most common method is to use a geoid model. Mathematicians have developed various geoidal models containing pre-determined offsets between MSL and an Ellipsoid at a predetermined grid interval. The best known geoid model is probably the Egm96 geoid, a worldwide model that was derived by observing the deviations in satellite orbits resulting from changes in gravity. The Egm96 geoid model have a claimed standard deviation of 0,10m per 10km`s. In exceptional cases sudden changes in gravity due to geological faults and concentrations can result in inaccuracies of up to 1 metre over as little as 10km`s.
Professor Charles Merry (from the University of Cape Town) developed a custom Geoid model for South Africa. This geoid model (called the Quassi Geoid Model) is based on the Egm96 model, draped over South Africa using a polynomial transformation calculated from actual ellipsoidal and MSL trig heights. The accuracy of the Quassi geoid model (i.e. the difference between geoid model separation and actual true separation) is much better than the Egm96 model, however, the precision is roughly the same because the sampling interval is the same.
Another method to relate ellipsoidal height to MSL is to do a GPS calibration. A GPS calibration is a mathematical transformation based on actual measurements using GPS on points with known MSL heights. The transformation applies a best fit plane between the surveyed beacons. The accuracy of the calibration is a function of the accuracy of the published MSL heights of the Trig beacons, and the accuracy of the GPS measurement. Although trig heights are published to the nearest decimetre, it is commonly believed that Trig height accuracies are no better than 0,2m in South Africa. The other problem with a GPS calibration is that a plane can very seldom accurately represent the Geoid, due to the uneven surface of the Geoid, except over small areas (approximately 10km`s x 10km`s.)
Examples
The following examples are based on typical accuracies achievable with the various methods, over a typical strip survey of 30 – 100 km`s in length.
| GROUND CONTROL METHOD | TYPICAL ABSOLUTE MSL ACCURACY FOR GROUND CONTROL |
| Single Base GPS Survey without Geoid modelling | 0,3m – 1,0m |
| Single Base GPS Survey with Geoid modelling | 0,2m – 0,5m |
| GPS Site Calibration | 0,1m – 0,3m |
| GPS Horizontal Survey with Double Run Spirit Levelling | 0,002m – 0,03m |
Conclusion
When high accuracy is required it is necessary to tie an Aerial or Lidar survey to an accurate ground control network. To achieve high accuracies in ground control survey it is necessary to use Spirit Levelling (for accurate MSL heights) combined with conventional or GPS equipment (for accurate horizontal positions.)
