Accuracy of UAV photogrammetry compared with Network RTK GPS
27 June 2013
Authors: Paudie Barry, Ross Coakley of Baseline Surveys Ltd
Baseline Surveys Ltd specialises in the supply of accurate geospatial data, such as cadastral, topographic and engineering survey data to commercial and government bodies. Baseline Surveys Ltd invested in aerial drone photogrammetric technology and needed to establish the accuracy of the geographic data derived from its unmanned aerial vehicle (UAV) photogrammetry before marketing its new aerial mapping service.
Having supplied the construction industry with survey data for more than 20 years, we felt that it was crucial for its clients to clearly understand the accuracy of photogrammetry so they can safely make informed spatial decisions, within the known accuracy limitations of its data. This information would also inform us on how and where UAV photogrammetry can be utilised.
The specific information required was the actual accuracy that could be reliably achieved using a UAV to collect data under field conditions throughout a typical 2Ha site. We flew a UAV over the test area in a typical matrix pattern with an 80% front and 80% side overlap; we placed 45 ground markers as checkpoints and surveyed them in using Network RTK GPS. We specifically designed the ground markers to meet our accuracy needs and established 10 separate control points – we inputted these into our photo-modelling software.
The rest of the GPS co-ordinated check marker data were added later in ArcMap to the completed orthomosaic and digital elevation model so that we could accurately compare the UAV photogrammetry XYZ data with the RTK GPS XYZ data at highly reliable common points. The accuracy achieved throughout the 45 checkpoints was 95% reliably within 41mm horizontally and 68mm vertically and with a 1cm ground sample distance.
This finding has shown that XYZ data derived from UAV photogrammetry has a similar practical accuracy to RTK GPS, which is commonly used for cadastral, topographic and engineering survey work. This means that UAV photogrammetry can replace GPS surveying as the main method of data capture for engineering projects, boundary mapping and topographical surveying. Aerial photogrammetry can now be used for projects with a 1:200 map scale accuracy requirement.
Much of current land, engineering and cadastral survey work using GPS and total station is often labour intensive; it sometimes involves surveyors working in hazardous environments; and the completeness of the data captured often depends on the time allotted to the survey project. No representation of geography is perfect and certainly the point, text, line and polygon style of 21st-century digital mapping is no exception.
GPS and total station geographic data collection methods are accurate enough to design civil engineering and architectural projects with sufficient practical accuracy (within +/-5cm) to design roads, bridges, pipelines and buildings.
Although GPS data is highly spatially accurate in both absolute and relative terms, it completely falls down in richness of data. Each line, point and polygon must be described by textual means in order to communicate its geographic meaning. Its completeness is something we do not dare even to consider doubting, but there is an undeniable risk of a surveyor omitting data without even realising.
This style of data collection is often time consuming and by nature expensive. We wanted to experiment with UAV photogrammetry to see if it would be accurate enough to replace current GPS and total station survey methods in engineering, cadastral and topographic surveys.
This could revolutionise the world of engineering surveying in terms of vastly reducing data capture cost, while vastly increasing data quality and richness.
As we were practically pushing beyond the current accuracy limitation of UAV-derived geographic data, we decided to use the C-Astral Bramor UAV for its superior stability as an aerial data capture platform. Its stability being achieved by a combination of its aerodynamics and high-end Lockheed Martin-designed autopilot system. The high 24MP precision of its Sony Nex-7 camera was another deciding factor, along with its very clever safety feature of a remotely deployable parachute landing option. It also has three-hour endurance and a practical wind tolerance of 20 knots when piloted by an experienced RPAS crew.
While we had decided on using Network RTK GPS to establish Irish transverse mercator co-ordinates on the centre positions of the ground control points, we carefully designed geometrically patterned ground markers which allowed us to obtain sub-pixel accuracy of the centre point location when identifying ground marker positions at the post data collection and pre-processing stage.
Agisoft Photoscan was the software we decided to use for its stated accuracy 1-3 pixels and the high quality of the ortho photography and digital elevation model outputs. ArcMap was used for presenting the final fusion of data captured by UAV and GPS.
The nine ground-control markers were theoretically positioned, so that they were equidistantly distributed throughout the site to ensure an even distribution of errors.
A flight plan was generated so that that there was an 80% overlap and a 80% sidelap and an altitude of 90m to provide an expected GSD of 10mm. Flight direction was plotted at 90 degrees to the actual wind direction, so as to maintain a constant ground speed of less than 16m/s during the photographic process. This step helped to reduce ground smear, a phenomenon that blurs the pixels due to the movement of the UAV.
DATA ACQUISITION SYSTEM
The data acquisition system we used on this project is a hybrid of Trimble GoeXR Network RTK GPS and a C-Astral Bramor UAV. We used the RTK GPS to establish ITM co-ordinates on our specifically designed ground markers to provide photo control. The GPS unit has a spatial accuracy in the region of 10-25mm both horizontally and vertically, due to the fact that we used struts to maintain steadiness during NRTK readings.
The C-Astral Bramor UAS platform is a blended wing, constructed of Kevlar and carbon fibre and has a 4kg MTOM. It is catapult launched, has extremely steady flight characteristics and advanced safety features afforded by its Lockheed Martin autopilot, including a parachute deployment system for emergency and routine landing procedures.
It carries a Sony Nex-7 24MP RGB sensor, which is oriented in portrait which allows for more forward overlap at a slower triggering interval. Our ground control markers were carefully designed so the most exact centre of the marker could be determined with a very high degree of accuracy from the photography at the processing stage.
As surveyors, we wanted to carry out our accuracy test in everyday conditions which, by their very nature, are sub optimal. The prevailing weather conditions on the day of our flight test were cloudy, with intermittent sunny spells. The wind speed at our flying altitude of 90m AGL was a maximum of 7m/s. The wind direction at altitude diverged from that used in our pre-flight planning by 20 degrees. This had the effect on the UAV of getting a 10m/s variation in its ground speed when travelling with and against the wind.
The topography of the site had a 5m variation in level, was surrounded by mature woodland up to 25m in height and had numerous buildings that would all contribute to turbulence at our flying altitude.
Firstly, ten control stations and 45 checkpoints, which were surveyed by Network RTK GPS in Irenet95 co-ordinates, were downloaded into Geosite office 5.1 and exported to AutoCADlt 2013 as two separate files.
We downloaded 1,601 photographs from the Bramor UAS along with the log file, which contains photo GPS position, barometric height, roll, pitch and yaw. We imported the photos and the logfile into Agisoft photoscan and, using the software, we eliminated superfluous photographs by deleting photos with high roll values, which occurred at turns.
The refined 728 photographs were then used for the photo alignment stage; the reason why we used an oversized area is that we felt that the additional photographs would contribute to alignment accuracy of the target area. We then imported the ten control stations into photoscan and identified the centre of each control point marker on each photo and attached it with its appropriate co-ordinate value.
Of the 728 photographs, we then used 168 photographs to further process the data into a 3D model for subsequent orthophoto and digital elevation model (DEM) output. The orthophoto was outputted at a resolution of 10.2mm pixel size and the DEM was outputted at 20mm GSD (see Fig 5).
The resulting ITM geo referenced orthophoto and DEM were imported into ArcGIS, along with the GPS CAD data. Distances were measured from the centre of the target, as they appeared on the orthophoto, to the centre of the target as measured by GPS by using the ArcGIS measure tool to attain the distance between both readings. GPS level point data were compared with DEM readings at the same point. Results were then recorded as shown in Figure 6.
The resulting geo-referenced orthophoto and DEM were imported into ArcMap along with the co-ordinated data from the RTK GPS. The difference between the two data sets is shown in Fig 6. The above results where imputed to Microstation to produce the graphs shown in Figures 8 and 9.
From the table shown in Fig 6, further interesting results were obtained (see Fig 10).
The significance of the ability of a UAV to capture geographic data at similar accuracies to RTK GPS along with the richness of detail that can only be revealed by 10mm GSD aerial photography is truly revolutionary. Like all disruptive technologies, UAV photogrammetry will take a couple of years to become mainstream, but when it does, it will almost fully replace current methods of engineering surveying. The accuracy that can be achieved by UAV photogrammetry is within 1:200 scales according to NSDI and FGDC mapping accuracy standards during sub-optimal conditions.
Now that UAS can be used to map at scales up to 1:200, UAVs will start to take over from GPS and total station as the main survey grade data collection method as UAVs are far more efficient at capturing geographic data, offering a huge financial advantage to government agencies and private companies using UAV technology to collect geographic data, particularly over more detailed and larger sites.
Apart from an enormous time saving on data collection without an appreciable loss in accuracy, UAV aerial photogrammetry offers far richer data than conventional survey vector data consisting of points, text and lines. Instead, UAV photogrammetry offers the user a bird’s-eye view of the site without any need for text or any fear of data being omitted.
The accuracy to which levels are generated by the photogrammetry allows for contouring at 0.2m intervals, which is very encouraging. GPS readings would still be required for manhole covers, finished floor levels etc, but even so, this still represents a huge leap forward in terms of surveying efficiency.
In terms of representing the landscape, the orthophoto can be combined with the DEM to produce very accurate photorealistic 3D modelling in programmes such as ArcScene and can be analysed to yield highly accurate earthmoving volumetric calculations. This is the most spatially accurate aerial survey data that we could currently find anywhere on the internet.
References are available on request
With a degree in civil engineering, Paudie Barry began his career as an engineering surveyor on thrust bore-tunnelling contracts on the London Water Ring Main project. He returned to Ireland in 1990 to set up Baseline Surveys Ltd at the age of 22. By 2007, Baseline Surveys had already carried out over 3,500 topographical, building and engineering surveys.
Barry has been invited into University College Cork’s Geography Department as a guest lecturer on the subject of drone-mapping technology. He is also a committee member of UVS International.
Ross Coakley BE, MIEI, worked in England as an engineer on numerous building projects before returning to Ireland in 1993. He worked as part of a design team designing water and sewage works before setting up an engineering consultancy in Bandon, Co Cork in 1996. Coakley specialises in house design using 3D CAD and simulation packages.
In 2012, he teamed up with Baseline Surveys to develop a UAV aerial photogrammetry GIS data-capture service. Baseline Surveys claims to produce the world’s most accurate aerial photogrammetry with a RMSE of only 4cm over a five-acre study area.