Over the past three and a half years Australian UAV has been capturing and creating data to inform waterway management through the provision of a cost effective alternative approach to ground based or LiDAR survey. Key amongst this data collection has been the delivery of high accuracy photogrammetric survey data for the assessment of erosion and to inform structural design.
The advantages of this method of data collection include the data capture turnaround to delivery time, accuracy, cost effectiveness and superior data-point density. This recent project for the Wimmera Catchment Management Authority created highly accurate survey data to inform the stability design.
Example outputs – 3D surface model and high-resolution mapping image
Eroding rivers are a feature of the Australian landscape, however, the clearing of land and grazing pressures increase the severity with the consequential release of sediment. This has and is continuing to have a detrimental impact on the ecology of rivers and receiving waters. Over the past 50+ years waterway and land managers nationwide have sought to address this erosion through the construction of amelioration measures, fencing and in some cases, sediment removal.
Aside from assessing the catchment sources of sediments and particle size, the first step to understanding a waterway and its erosion and sediment characteristics is often to examine the potential for scour, transfer and deposition of sediment, for this there is a need for detailed and accurate survey.
The traditional specification of a ground based survey of an eroding waterway would be to capture the top of the bank and thalweg with cross-sections at areas of significant change or at 100m intervals. From this data the longitudinal gradient and channel flow velocity would be modelled in one dimension giving an indication of scour potential and sediment movement. Traditional surveys are very accurate where the point is taken but may overlook important features in-between leading to broad assumptions being made. These ground based surveys are time consuming, expensive and increase the safety risks of slips, trips falls and snake encounters for those undertaking them.
The advent of LiDAR greatly enhanced the ability to model the flow of water in channels. Having a three dimensional points throughout the whole channel created a more complete picture. The spacing of these points are typically between one to four points per square metre with a vertical accuracy of plus or minus 15-20cm. Where LiDAR has previously been flown this is a good option for modelling but there is still a requirement for supplementary survey for detailed design of instream structures. If LiDAR is to be commissioned on a reach scale it is typically cost prohibitive. Also, given the dynamic nature eroding of channels, the time elapsed between capture LiDAR and use of data may also compromise its usefulness.
The Right Tool for the Job
When choosing a UAV for data capture it is essential to assess the key constraints and scope of the project. While much of the media and hype is focused on multi-rotor drones, for this type of work they are compromised in the area they can cover efficiently making the project more costly than it needs to be. Because we specialize in this type of large area rural capture work our team has developed a fleet of cost effective UAVs that can capture up to 1000ha a day at 2.5cm pixel resolution, this exceeds the capabilities of all but the most expensive commercially available UAVs. This high image resolution is obviously important for the identification of features in the images but also translates to accuracy of the 3D models.
AUAV’s custom-built UAVs and an automated AeroPoint survey target
‘Survey Grade’ Versus ‘Mapping Grade’ Accuracy
A typical UAV GPS is accurate to around 5 metres in the horizontal and far less accurate in the vertical so the resultant three dimensional model will always be too inaccurate to be useful in modelling and design. For this project the way to address this inaccuracy was the placement of ground control. These points on the ground are recorded with a high accuracy GPS or, as was the case on this recent erosion project, Propeller Aeropoints are left in place and corrected to be highly accurate in post processing. These known locations are then used in the processing to anchor the resultant model to a known elevation which in turn creates the highly accurate ground model.
The spacing and locating of these points will also directly affect the accuracy of the model, if they are too far apart there can be inaccuracies introduced into outputs, each project is planned to ensure sufficient control is placed to gain an accurate result that is fit for purpose.
With one of these two methods the fluctuations in the accuracy in the horizontal and vertical are reduced to the order of 2-3 x the pixel size or better. Therefore, in this recent example project, there is a pixel point on the ground every 2.5cm and that point is accurate to ~5-7cm in the vertical. This compares to the manned aircraft LiDAR of between one and four points on the ground per square metre and a vertical accuracy of 15-20cm. The advantages of the additional data are apparent when comparing the datasets.
We have found that every client has a different data need and ability to manage large datasets. Keeping this in mind the data that comes out of the photogrammetric software package is rarely useful in its raw format for the client due to the size of the data and therefore needs additional work. For this reason we generally thin the points on the flatter ground while maintaining the higher point density in the areas where detail is required. This reduces file size and data usability.
The images at 2.5cm/pixel are clear enough to have a good clear picture at 1:100 scale in GIS which is then delivered in a compression format to reduce size ensuring it is compatible with the software the client is using.
Zoomed in to creekbed view at approximately 1:100 scale
On occasion we are also requested to complete terrain analysis or generate additional outputs such as contours, top of bank, cross-sections and long-sections. It is always better to have too much data and have to discard some than to not have enough and make assumptions when undertaking this work. With the combination of the high resolution images and accurate terrain the delineation of features in GIS or other design packages becomes straightforward.
Stream Bed Profile Analysis
As a company we also work with a range of other consultants that can undertake the hydraulic modelling, rainfall-on-grid hydrology, scour assessments and design services as required.
Relying of photogrammetry does have some limitations. In areas of dense vegetation the technique will not provide a good result. Aerial LiDAR is better at penetrating vegetation and therefore might be the better option. Our team will always advise if this is the case and we work with LiDAR capture companies and can facilitate this. While reliable, affordable and accurate UAV mounted LiDAR will come in time, for now though the default position in densely vegetated areas is manned LiDAR with UAV mounted LiDAR still very costly and in the realm of experimentation in the research organisations.
Neither photogrammetry nor LiDAR will work in rivers full of water. In this circumstance a great option is to call upon the bathymetric survey boat to work in tandem with the UAV. The boat and UAV datasets stitch together readily to provide a complete picture. We can also assist in coordinating the capture of this data.
Bathymetric survey boat
In addition to waterway erosion surveys Australian UAV has applied the same methods to survey:
- Cliff recession
- River diversion design
- Mine subsidence
- Irrigation channels
- Dune and beach erosion.
If you have any questions regarding the use of UAV data in the assessment of waterway erosion and analysis please don’t hesitate to drop us a line at 1300 738 521, or email firstname.lastname@example.org
James Rennie worked in the natural resource management industry specializing in river and water management for 17 years prior to founding Australian UAV.
Copyright Australian UAV 2017 (www.auav.com.au)