Terrain Modelling

Digital Terrain Model (DTM)

Bare earth model elevation, excluding all non-ground points such as built (bridges, roads, towers) and natural (trees, vegetation). 

Lidar

Land-use planning, infrastructural project management, soil science, hydrology and flow-direction studies

Digital Surface Model (DSM)

Capture of all natural and built features (including tree canopy, vegetation, buildings and powerlines).

Lidar or Photogrammetry

Urban planning, infrastructure inspection, aviation planning, vegetation management, visual line of site.

Canopy Height Model (CHM)

An extracted model taking the difference from DSM and DTM.

Lidar

Forestry Management.

The relationship between the models is best represented as follows:



LiDAR (Light Ranging & Detection)

LiDAR is pulsed light which bounces off the land or objects used to gain a range (distance). The known source position (using Inertial Measurement System and Global Positioning Systems) of the laser light pulse allows triangulation. Surveyed ground control markers and GPS base-stations are used to verify corrections including that for national grid. The aggregated light pulses create a very large point cloud of position and elevation values (XYZ). The greater the LiDAR source positioning, strength and frequency, the better the light penetration through tree canopy and vegetation and better outcome models (accuracy and reliability). LiDAR surveys are therefore particularly useful for all DTM, DSM and CHM. Paired imaging systems can enhance true view texture for more advanced visualisations.

Pros: Exceptional at obtaining high level of land coverage including penetration through trees and vegetation. (subject to Lidar strength and INS). Solution works in homogenous environments and varying light conditions and night.

Cons: More expensive solution / Weather: susceptible to fog and rain 

Photogrammetry

Photogrammetry is the science of making measurements from photography. The process uses a process of triangulation but in respect to photo images (rather than points of light reflectance, as the LiDAR). The survey platform is therefore required to be from a known position, which is achieved using global positioning systems. The overlapping images are then stitched together using recognisable and similar markers, which can include surveyed-in ground control marker points. The ground control points are then used to support the corrections required to achieve a national grid based orthorectified mosaic (orthomosaic) combined image. The process can also be used to develop a point cloud (individual data points with xyz coordinates) which can be used to support generation of a DTM and a DSM vegetation permitting.

Pros: Relatively low cost solution in obtaining basic DSM mapping (subject to camera resolution, height of survey and ground control).

Cons: Weather and Light: Susceptible to rain, fog, wind and poor/highly variable light conditions / Homogenous Surface Areas: Low contrast and uniformly textured surfaces such as sand, water, short grass can provide difficulties in the software matching processes resulting in increased errors.

Survey Processes

The survey approach will always typically involve significant planning and preparation of the subject area to be surveyed. This will include assessment of the area to ensure that it is safe to fly and all necessary permissions are in place prior to any site work.

Survey teams (typically a pilot and a spotter) will base themselves in key positions to maintain Visual Line of Site (VLOS). The area of survey can be extending by moving the base or applying additional spotters for Extended Visual Line of Site (EVLOS). Larger areas, where VLOS and EVLOS are not possible, require a specific safety case for approvals for Beyond Visual Line of Site (BVLOS).

Although many drones are fitted with Real Time or Post Processing Kinetics (RTK/PPK - a process to enhance positioning using global navigational satellites in real time or post flight) it is best practice to ground truth all survey based data from drones. This is achieved using Ground Control Points (surveyed ground based marker positions) and with RTK base station(s).  Ground Control Points are essential for supporting surveys which are to be repeated such as peatland restoration, mining or coastal surveys.

What survey method should I select?

The survey choice is not necessarily an either/or, but with advances in technology both methods can be applied, and at the same time. This offers the best of both worlds with far richer datasets. However, as a guide the core decision making factors include:

LiDAR

Mapping of difficult to access, complex and overgrown terrain

Capturing details on thin, vertical or man-made structures

Projects where details and precision are priority

Where detailed vegetation structural diversity and modelling is required

When there is low light or varying light conditions.

Photogrammetry

Context rich scanning which requires minimal post processing and expertise.

Datasets that need visual assessment.

What should I consider when procuring a survey?

Not all UAS Platforms, sensors and workflow are made equally. You could potentially achieve a survey using a low cost off-the-shelf drone and have some nice images but then no reliable usable data. Robust data is the foundation for all of your analysis, future monitoring and your evidence. Key things to consider are:

Objectives: What you wish to achieve (basic terrain, volumetric, biomass, modelling, analysis, structural, image)

Size: The size of an area will determine the survey method, height and sensor grades used.

Accuracy/Resolution: The greater the accuracy the larger the dataset, the lower the flight and the higher sensor rating. A photogrammetry survey could be achieved with a basic 20mp camera flying at 40m to achieve a <10cm RMSE or a 61mp camera flying at 80m to achieve a <2cm RMSE. A lidar survey can use a basic sensor (solid state) achieving 300,000pts/s through 8 channels at <40m and give <10cm RMSE or a higher end sensor at 1million pts/s across 32 channels at <80m to achieve <2.5cm RMSE. Be clear on what is required and don't over specify.

Analytics: Photogrammetry and low end lidar are usually not sufficient for more advanced assessment (volumetric temporal changes, vegetation structural analysis, biomass etc).

Timing: Lidar can be used any time of the day or evening, while photogrammetry is impacted by time of day. Quality useful images avoid shadowing effects which result from sun position and weather changes.

Data Size: The raw and processed datasets can increase in size dramatically when higher resolutions are specified. A typical outcome from a higher quality lidar and photogrammetry survey across 400ha will result in 4Tb of data.

Flight Management: Good surveys are fully compliant, well planned, insured, HSE conscious, involve engagement with land/property owners and take privacy seriously.

Further factors to consider are weather, land homogeneity, data sharing, environmental/ecological sensitivity, flight hazards and seasonality.

Applications

There are a range of applications where LiDAR and/or Photogrammetry can be applied. These can be enhanced when used with Multispectral, Thermal or SAR sensing. Sensing can be achieved at scale from a range of platforms.