Drones: Blowing away the competition in assessing short term Coastal change

Why drones?

Drones provide imagery with high temporal and spatial resolution, and the fact that they are low cost, as well as easy to deploy and use, makes them one of the best platforms to study short-term coastal change. Many researchers have noted that the drones are more accurate, and can detect subtler changes than other techniques. Drones also allow morphological beach change to be studied in conjunction with shoreline change, as satellite imagery can be too coarse to pick up the former, and LiDAR cannot easily identify the land-water boundary.

But how to drones detect coastal change?

Most remotely controlled systems have both optical and infrared sensors attached to them. The optical sensor employs wavelengths in the visible range of the electromagnetic spectrum (0.4-0.7μm), and the output imagery is usually a red/green/blue (RGB) composite, but can also be in greyscale. Tonal variations can be used to delineate the shoreline/cliff location as well as the dune position, however RGB imagery can be more accurate, as it allows better visualisation of shadowing. The infrared sensor uses the near infrared range of the spectrum (0.75-1μm), and is especially useful in locating the shoreline as water characteristically absorbs highly in the infrared portion of the electromagnetic spectrum. The raw imagery (figure 1 [1]) goes though some quality checks, and then an algorithm called structure from motion (SfM) produces calibration measurements and a 3D point cloud by identifying feature points in the data, and assessing how these move throughout the imagery. The 3D point cloud can be used to create a dense point cloud (figure 1 [2]) and then a digital elevation model (DEM) that is georeferenced with ground control points (GCP’s). Another algorithm called MVS is needed to assess changes over time, which involves combining the dense point cloud and producing 3D content from overlapping images, DEM’s (figure 1 [3]) can then be georeferenced and analysed in a mapping software, such as ArcGIS.

Screen Shot 2016-05-28 at 2.26.33 PM
Figure 1. The process outputs of converting raw drone imagery [1] to a dense point cloud [2] and then finally to a digital elevation model [3] (Casella et al. 2016)
Once the DEM’s are produced, then these can be combined together with others from different flights (times), they then can be used to assess a number of coastal changes such as; beach erosion/accretion, cliff retreat, dune stability/succession, shoreline variation, and be used in various predictive models. An example of shoreline movement and dune variation can be seen in figure 2. However the analysis outputs cannot be interpreted without knowledge of many natural variables that are known to impact the coastal environments and these range from sediment grain size/coarseness, and wave/wind/tide variables to large scale current flows and climatic influences such as ENSO.

DEM
Figure 2. Example of an output from drone imagery that is analysing beach change from Nov-Dec 2013 to Dec 2013-Mar 2014 and shoreline movement (Casella et al 2016).

What about the other methods of assessing coastal change?

There are two other main methods of assessing coastal change, and these are (1) using satellite imagery and (2) LiDAR. GPS can also be used to log the position of specific features, but cannot be visualised without the use of imagery, and is therefore used in the majority of studies as GCP’s in the georeferencing process.

Satellites such as Landsat and IRS have spatial/temporal resolutions of 16 days/30m and 24 days/23.5m respectively, and have a much larger range of sensors on board. The resolution of the imagery received is well suited if you are looking at long-term change over large areas, of if you want to look at dune vegetation. However you cannot choose when images are taken, and therefore won’t be able to assess how the beach responds to short term pulse events (such as storms), and the spatial coarseness of the imagery may lead to inaccuracies at meter scales, or entirely miss dune systems.

LiDAR on the other hand, uses an active sensor (lasers) to provide a range measurement to a point on the ground from a single airborne position. It is more accurate than satellites (1-2m), but the temporal resolution of the scans can be limited, does not date back very far in time, and they are a lot more expensive to run than drones. The limited amount of past imagery can reduce its usefulness for assessing long-term changes, and the temporal resolution/cost involved would make it difficult to monitor pulse events.

A holistic view of coastal change

Drones obviously have their downsides, they cannot carry many sensors, have difficulty in assessing large areas, may be difficult to fly under certain conditions, and might become damaged from the sea spray. Drone technology, like all other technology will continue to improve in the future, resulting in sturdier platforms (more resistant to the coastal environment) that can hold more sensors, for a longer period of time. This will continue to improve the output and therefore applicability of drones in studying coastal change and informing coastal management. However we are still missing data that suitably describes the causality of change in the coastal environment, and there has been limited evidence of studies that combines remote sensing platforms. This is in itself limiting the assessment and prediction of coastal change as coastal environments, like many natural environments, are connected to forcings at multiple scales. If we were to analyse the coastal environment using different platforms that have varying spectral, spatial, and temporal resolutions then it is likely that we could make better, more informed decisions within coastal management.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s