UAV vs SfM: What Is Flight Planning in Photogrammetry?
Aerial photogrammetry is at the forefront of a research and development revolution right now, thanks in large part to the expanded use of UAVs.
More commonly known as drones, these unmanned flight solutions have lowered the bar of entry for collecting aerial images, making the technology cheaper and easier to access for more researchers. The result has been an explosion in innovative new use cases in industries like telecommunications, environmental science, oil & gas, and more.
Perhaps most importantly, new flight technologies have streamlined flight planning for photogrammetry image collection. Flight planning has been an important aspect of aerial photography since the earliest airborne recon flights a century ago.
In this post we’ll go over how to create a flight plan for collecting photogrammetric images, as well as the difference between Structure from Motion (SfM) and traditional UAV flight plans. We’ll also dig deep into the all important question: what is flight planning in photogrammetry (and why does it matter?)
What are flight plans and why are they important?
To get a high-quality orthomosaic map, you need high-quality images. To get high-quality images, you need to plan for a number of different variables, and develop a flight plan that creates the right conditions to handle them.
Without a defined and catalogued flight plan, even the most detailed aerial images lack important context needed to build an orthomosaic composite that is reliable, detailed, and accurate. That’s where flight planning comes in.
A drone flight plan will ideally be uniform in every possible sense, from elevation above a target object or landmass to airspeed and camera angle. Automation helps make the physical maneuvering of the drone easier, but without a carefully researched and detailed flight plan, the images you capture will end up with distortion that undermines accuracy and usability of the final map.
Some important considerations that go into a flight plan include:
- Flying height above a subject (not the altitude from sea level)
- Size of the sensor (a.k.a. a charged coupled device or CCD)
- Image resolution in square meters per pixel
- Dimensions and shape of the subject area
- Ground speed
- Data related to number and type of laps
- And other data points as needed
Flight planners should anticipate around 70% image overlap to prevent gaps, distortion, and vignetting in the final map. Some projects may require more or less overlap depending on the nature of the research performed.
All of this information should match flight notes collected in the metadata of images for later processing. The metadata should also document who created a data set and the environmental conditions it was created under, including weather, humidity, and so on.
Depending on the type of photogrammetry software you’re using, you can either flight plan for traditional photogrammetry or Structure from Motion (SfM). Here’s the breakdown of both and what makes them unique.
How does traditional UAV image capture work?
To perform traditional UAV photogrammetry capture, a technician establishes a geometric flight pattern over the target area. This grid pattern will allow for overlap as required by the project guidelines. Images are taken sequentially along established flight lines, and images created during take-off, landing, and turns are typically discarded.
These high quality images, known in the industry as orthophotos or orthoimages, are pinned using GPS positioning with data on the exact positions and conditions of their creation.
Once collected, orthophotos are normalized using the pinned environmental data and organized into data sets that are sent to servers for collection, storage, and processing.
The collection process is technical, but precision and very careful planning are equally important for generating good results. Having the field experience to recognize features like ground control points (GCPs) during planning can streamline processing and avoid distortion by centering the flight plan on touchpoints that the mapping software relies on for calibration.
It helps to work with established professionals who are well versed in the most up-to-date photogrammetry techniques and tools.
What is SfM and how is it used?
New on the scene is SfM, a technique that introduces depth into calibration and modeling of two-dimensional images. The result is a reconstruction of a three-dimensional scene or object with multi-facing perspective and angles, just like the real thing. SfM technology can be used to create incredible 3D renderings from point-cloud image data that rival LiDAR in their measurability and multi-planar proportionality.
Using a SfM model allows for more idiosyncrasies in the flight plan. This technology relies on point-cloud measurements rather than GCPs, so it’s accommodating for variations in elevation, camera angle, and more.
This makes SfM a particularly good approach for natural resource and geospatial applications — for example, determining tree biomass or analyzing geology — and for situations where generating defined GCPs proves difficult.
Building a flight plan for SfM image capture may seem less rigid, but it requires no less preparation or rigor. If you want to use the UAV to document multiple sides of a construction project or high-value infrastructure, you’ll have to plan for variable camera angles and heights. You’ll also want to work with an experienced provider that has field knowledge about this innovative new toolkit.
The importance of flight planning for photogrammetry projects can’t be overlooked. Whether you’re prioritizing sharp imagery or 3D functionality in your orthomosaic maps, the quality of the final product will be a testament to the preparations you make.
And no matter which approach you use, you need powerful photogrammetry software to help create an orthomosaic map from your data. Mapware from Aerial Applications is a cloud-based orthomosaic mapping tool that is powerful enough to produce large, detailed 3D orthomosaic maps quickly — and user-friendly enough to make the process enjoyable.