The purpose of the data was to provide natural surface levels to support the design of a new mains water pipeline from Normanville to Wirrina Cove. The total capture area was 530 hectares.
Challenges included steep and variable terrain (ranging from 0 to 140m within a single flight corridor), including cliff areas, windy conditions, and visibility limitations over steep rolling hills.
Long grasses and dense tree canopy across paddocks made LiDAR a more suitable solution than photogrammetric modelling.
The total area covered was 530 hectares across three mission areas and five flights. The maximum area covered per full battery varied depending on wind conditions and mission geometry. Generally, when flying at 110m altitude, up to approximately 165 hectares were
captured over a 35-minute flight.
Due to the undulating terrain, mission settings were configured to AGL with Real-Time Follow enabled, which effectively maintained the required altitude above ground level. Flight lines were aligned parallel to cliff faces to minimise sudden elevation changes along the flight path.
Mission Settings Summary:
Payload Settings:
These payload settings were selected to optimise the accuracy of processed LiDAR data, particularly on road surfaces.
For positioning, SmartNet was utilised, as the nearby CORS base (5YAN Yankalilla) was within 10km of the furthest mission area. RTK Service Type: Custom Network RTK.
Approximately 12 Ground Control Points (GCPs) were established within each mission area using internet-corrected GNSS. These points were evenly distributed throughout the model and selected on smooth, regular surfaces—typically bitumen.
Sprayed white crosses on bitumen were also used during post-processing to verify X, Y, Z conformance. Traditional survey teams collected check shots across roads (“QQ checks”), which were processed in DJI Terra as checkpoints.
Important Note: Ensure the distance to the CORS base does not exceed 15km for any mission area. While it is possible to fly beyond this range with flight lines appearing correctly georeferenced, reconstruction in DJI Terra may fail.
Final GCP accuracy was generally within 30mm, with the vast majority of check points falling within 40mm. A small number of outliers were observed, with height deviations of up to 80mm.
With a 16-point return density, LiDAR provided excellent coverage beneath tree canopies and through dense low vegetation—something not achievable using photogrammetric methods alone.
The Zenmuse L3 also delivered significant efficiency gains through its dual 100MP oblique cameras, enabling simultaneous capture of LiDAR and ortho-imagery. Reduced overlap requirements (40% side overlap compared to the traditional 70%) resulted in substantial time
savings. Each mission only needed to be flown once to generate multiple deliverables.
In summary, the Zenmuse L3’s ability to combine LiDAR and ortho-imagery in a single flight delivered significant time savings compared to traditional methods.
The Matrice 400 demonstrated excellent stability and was capable of operating in winds up to 12 m/s, providing confidence in challenging conditions.
Additionally, the Matrice 400 offered safety advantages over fixed-wing platforms, with highly accurate and repeatable flight paths. The full-colour Field of View (FOV) cameras provide real-time on-demand visual coverage to the front, rear, left, right, and downward
directions, enhancing situational awareness when operating over sensitive areas.
