The latest in the Drones and Small Unmanned Aerial Systems Special Series, in which Kike profiles interesting information, research and thoughts on using drones, UAVs and remotely piloted vehicles for journalism and photography.
The combination of light (photo), drawings (gram) and measurements (metry) are known as photogrammetry. Until recently, photogrammetry was a very specific niche within the geospatial industry: a rather complex subject, accessible only to specialized professionals. However, this field is becoming more and more broadly accessible. Photogrammetry has become a field in which, despite the need for some initial training and effort to understand the basics, there are many useful applications such as creating 3D maps using aerial images.
To put it simply, aerial photogrammetry is the act of mounting a camera on an aerial vehicle and taking multiple photos of the ground as it flies. The resulting visuals are then processed in a powerful computer using specialized software that creates a survey-grade output that is georeferenced and scaled. This allows us to make measurements from photographs, giving photographers the chance to pinpoint exact positions of surface points. With the advent and continuous evolution of computer imaging, the process of photogrammetry is becoming easier and easier.
Photogrammetric platforms (manned planes) are still carrying heavy, sophisticated, expensive, and accurate sensors. If you use such systems, the need for powerful software to resolve photogrammetric problems is less relevant. The emergence of Unmanned Aerial Vehicle (UAV) photogrammetry has been a game changer. Suddenly, light, low cost, low accuracy sensors without a gimbal have become airborne. Traditional photogrammetry had no good solution for how to handle such data. New generation software, such as Pix4Dmapper, has contributed substantially to this concept of “simply mounting a camera on an aerial vehicle and taking multiple photos of the ground as it flies”.
The camera is commonly airborne with its axis vertical, but oblique and horizontal (ground-based) photographs are also possible. Data reduction, converting data into a more usable dataset, is accomplished by stereoscopic line-of-sight using both analytical and analog methods. In vertical aerial surveys adjacent photos are overlapped. The two images of the same terrain are then superimposed for three-dimensional viewing by human operators or automated sensors.
There is also something called close-range photogrammetry (CRP). We will not dedicate time to this, but in essence, it consists of positioning everyday cameras closer to a subject to measure buildings, recreate forensic scenes, or model archaeological objects. It is also referred to as “image-based modeling.” There used to be real differences that justified the different naming between CRP and aerial photogrammetry, but nowadays, both techniques are 99% the same.
With UAVs, everyday cameras, and new software available, the analog (line of sight) geometry is now obsolete. By analyzing almost every pixel in the images, and comparing highly overlapping sets of images, software packages are able to compute back to the 3rd dimension, in the same way the human brain gets a sense of depth from the two eyes.
With these images, scientists (or even you, if you learn how to do it) can create numeric models of natural systems. This creates the possibility of developing predictions and controlling the effects of natural disasters. From earthquakes, to tsunamis, to the flow of rivers descending from a mountain, the forces of nature can be more easily measured.
Note: This article is an excerpt from the new book So You Want to Create Maps Using Drones?