The aerial survey of Gorongosa National Park was one of the first projects where the firms new ELMAP laser scanner and multi-head camera system were both deployed. Developed at WGS and designed to offer the same across-track footprint as the ELMAP, the multiheaded camera system perfectly compliments the ELMAP and results in higher survey efficiency. Logistics and weather played along well and made way for three perfect days of flying.
The image was captured at 15cm/px and the LiDAR was used to create a 50cm digital terrain model. LiDAR data was used as the input for a machine learning algorithm aimed at counting termite mounds which are all but invisible in the orthophoto but clearly visible in a colourised height model.
The largest single survey undertaken by WGS occurred over the course of three weeks in the Canadian province of Saskatchewan. Very strong winds that came from any direction on any particular day and low clouds made the job tough to plan. We used multiple versions of flight lines for the differing wind directions. When the low cloud did not obscure the survey site, or a large enough potion of it, the WGS team were airborne at first light and spent long gruelling days roaring over the quietest grassland ecosystem on Earth.
There were a number of challenges faced by the team during acquisition but the real work came with processing hundreds of terabytes of data, captured over multiple days. The operation accrued so many flying hours that the aircraft reached its service interval midway through the survey and the equipment had to be removed and reinstalled to make space for the 100 hour service. The resulting spatial information has been used to identify and quantify the location, size and volume of “prairie potholes” or “kettle ponds” formed by retreating glaciers. These kettle ponds fill with water outside of winter months and cannot support agriculture until effective drainage is applied or they can be filled in.
UAV systems are a challenging photogrammetric problem. It is just not yet possible to install survey grade instruments to a small drone. The point and shoot style cameras provide a challenging platform to work off due to the low cost and overall diminished precision that the instrument is manufactured to. The portability and lower precision of drones makes them perfectly suited, however, to small sites that require close range survey with a very small scale.
Manilla required such a survey for the purposes of mapping all visible terminal structures and utilities as well as every paint marking. The very small scale that features were mapped to provide the basis for automatic container logistics. WGS processed the thousands of small-scale images and built a seamless orthophoto with a 1cm GSD which allowed the staff to capture even the smallest painted text in the container terminal. Ultimately the supplied CAD drawing serves to provide a map for machines to navigate the terminal and keep track of where they, and the containers are, at all times.
WGS flies a number of smaller towns throughout the Canadian season, providing 10cm orthophoto and high density, airborne laser scans that are used, among other things, for town planning and infrastructure observation.
The sensor combination that is used for LiDAR and Orthophoto allows WGS to acquire both in a single flight, maximising efficiency and lowering costs.
WGS flies areas included in the Alberta Forest Management Agreements for long-term, renewable forestry on a yearly basis using multispectral sensors which capture additional wavelengths of light. The sensor captures infrared light in addition to visible bands, allowing various vegetation indices to be calculated which provide the forestry industry valuable insight into the overall health of the timber.
WGS uses a large format, multispectral Vexcel Ultracam to maximise the efficiency of the survey. The Ultracam is designed to provide a very large footprint yet still capture very high-resolution images. This allows the survey aircraft to cover a larger area from a higher altitude at the same resolution.
A purpose-built oblique imaging, multicamera rig was recently developed at WGS. The system is comprised of five cameras, synchronised and linked to survey grade GNSS and inertial measurement systems. Each camera captures a different orientation: forward, right, back and left facing 50mm
cameras point 23° below the horizon and a single 24mm camera faces straight down. With the correct flight planning all faces of a particular façade will be covered by multiple frames, from multiple angles, resulting in a robust object geometry.
The result is a digital twin of the area flown. A very high-density point cloud is derived using photogrammetric techniques combined with high end computing hardware. This point cloud is used to generate a geometric model of the city. A texture is aggregated from the input imagery for each model and draped of the triangulated faces to produce a photorealistic model. This model is mapped to real world coordinates in all three dimensions which allows every surface to be measured. Road width, heights of streetlights, angle of roof sloop and viewsheds can all be measured. With this exciting new technology, entire cities can be digitally hosted online and be fully navigable, with every aspect of the city capable of being measured, recorded and catalogued digitally from any computer with an internet connection.
Modular system design allows WGS to mount combinations of sensors and maximise efficiency. One such example was a survey over an industrial effluent site where RGB, IR and Thermal imaging sensors were all used concurrently in a single flight. Careful system design, calibration and operation
ensures that each sensor may be ultimately combined to provide specific measurements for the same point on earth.
Each sensor’s imagery is processed independently and combined in a final step to ensure consistency between datasets. The thermal imagery was further resampled using bicubic interpolation to match the 10cm/px resolution of the RGB ortho and the two were combined to produce an augmented RGB orthophoto.
Satellite sensors provide coarse image data with additional image bands beyond just RGB. Satellite data can be acquired at fairly regular intervals, depending on cloud cover, making comparative analysis possible. Modern satellite services are offering data at continually higher resolutions and lower costs as the technology evolves.
WGS has been able to utilise freely available Sentinel II imagery to analyse change over time and provide confirmation of harvest data. Using existing field polygons as a vector layer, an average crop health indicator value for all raster cells within each vector polygon was determined. This value is
automatically calculated each week if cloud free satellite data is available and compared to the previous value. This allows WGS to analyse hundreds of square kilometres of cropland completely autonomously and identify when a field has been harvested within a week of harvest.