Wynand Uys, in the NE of South Africa, from a town called Hoedspruit. I’ll be using photographs to illustrate my introduction because the environment we operate in is quite different to what most community members are familiar with.
We operate a fleet of hot air balloons (for paying passengers) and two fixed-wing aircraft. No, not fixed-wing drones but ‘real’ aircraft that people sit in! Sorry, I had to include that little dig aimed at the drone industry. Nowadays, when you present aerial survey projects, people assume that a drone was used.
We got involved in aerial photography because part of our work is to provide air support to the folks who protect rhino in the nature reserves that are associated with Kruger National Park.
As an added capability to simple, human eyeball surveillance, our involvement in photography started with batches of oblique or nadir images to record events and locations of animals, people, and cars.
Driven by the conservation industry’s demand for information on the whereabouts and movements of animals, their numbers and the state of vegetation, we became formally involved in photographic census of large mammals, particularly rhino, buffalo and elephant.
The more products we delivered, the more we were encouraged to develop new kinds of projects. One of these projects is to annually determine sex and age ratios within each population (or herd) with really high resolution images.
Needing ever larger composite images of large herds of animals, I (naively) considered photogrammetry as a possible solution. Of course this did not work because of the dynamic nature of the scenes. Unless we could gather all the images for a single composite in a matter of seconds, there was too much movement of subjects and even if the ortho was successfully constructed, there were headless animals and some with two heads and also ghostly, transparent duplicates of animals. We adapted our photo-gathering techniques and for live animal studies we now only use panoramic stitching.
BUT: By fooling around with Pix4D in 2014 (initially for the wrong purpose), we stumbled across new products that were immediately in demand. Terrain modelling, vegetation mapping, demographic studies, agriculture, forestry management, land rehabilitation and more. Now we’re quite heavily invested (for a small business) in two aircraft, three pilots, substantial computing power, three High-end DSLR’s with pretty good lenses and a MicaSense RedEdge. I was very pleased, about two years ago, when Pix4D offered me the permanent license at a price discounted by all the monthly subs that I had previously paid. Photogrammetry has become an essential tool for us and I am pleased with the ever increasing capabilities contained in the updates.
Although I realise that Pix4D evolution is driven by the drone industry more than anything else, I am pleased that it works well with piloted fixed-wing aircraft.
We may suffer the disadvantage of a minimum high speed of about 65kt (120km/h) and a more costly platform, but high speed and no altitude restrictions mean we can survey ten’s of thousands of hectares in a day. We can stay airborne for 6 hours non-stop (even longer, but that leads to fatigue). That means we can fly to a survey area (cheaper than driving there), complete a large chunk of work, usually all of it, and return home on the same day without landing.
With the fast shutter speeds now available in high-end DSLR’s we find that aircraft speed is not much of a limitation anymore. Although most of our projects are thousands of hectares in extent, some even hundreds of thousands of hectares at GSD 15 - 30cm/pixel we can also manage smaller projects at 4cm/pixel. That’s about as high-res as we can go unless we use a slower aircraft or a camera intervalometer that can fire more rapidly than 1 activation/second.
Two days ago, we ventured into archaeology for the first time. We departed from our airfield at 9am, arrived at the remote and relatively inaccessible project area 200km away from base just before 10am, surveyed 2 sq. km at 4cm/pixel and arrived back at our hangar by 11:30. Pix4D ran overnight and I have, in front of me a beautiful model of ancient ruins and surrounding landscape.
This project would have been well suited for a drone survey, but the client was unsuccessful in getting clearance for a drone flight inside the Kruger National Park at short notice. Thank to our existing involvement as an air service in the Kruger Park we got instant clearance and the project is in the bag.
We have been so busy with protected wildlife areas that we have yet to develop agricultural products. The customers are waiting and the MicaSense Rededge is gathering dust.
My wish is that Pix4D continues with its fantastic work, developing the capabilities that support the drone industry without forgetting that the software is also used by piloted aircraft and very large projects, often with as many as 10 000 images.