Making photogrammetry more accessible and relevant
Our relationship with photography has evolved dramatically over the last few decades. What was once a relatively scarce skill, has never been more accessible and relevant to people’s lives. For example the selfie has come to represent a pillar of our culture, and yet as a technique, it remained out of reach for most people until the 21st century.
While we may understand photography as a tool of history, that helps rekindle our memories, and frame our past, the role of photography for data collection is relatively recent. As machine learning and AI technologies improve, machines are increasingly able to interpret and analyze photographs, at a rapidly increasingly scale and complexity.
The dronie, or drone based selfie, is not yet ready to replace the selfie as the dominant form of self-expression, however drones are enabling a dramatic expansion and advancement in the field of photogrammetry. The science and technology of obtaining reliable information about physical objects and the environment, photogrammetry is the process of recording, measuring and interpreting photographic images and patterns of electromagnetic radiant imagery and other phenomena (quoted from Wikipedia).
A major application of photogrammetry is creating 3D images or models using 2D photos. Another is to create accurate or realistic maps. Or panoramas. Or analyze crop fields or vegetation.
Photogrammetry is one potential answer to the question: “what are you going to do with that drone?”
Fortunately for us, there’s also a “future tool” that makes photogrammetry accessible, and it’s the focus of today’s issue.
Open Drone Map, or ODM, is an open source photogrammetry toolkit that can process aerial imagery into maps, 3D models, and similar composite imagery.
As a toolkit comprised of free and open source software, ODM is designed for professionals and researchers, but can also be used relatively easily by anyone. Like other free and open source projects, the difference comes down to depth and level of customization.
For example, while the software is entirely free, one of the ways the project raises money to fund development is by selling installers for the software. While you can go through the trouble of learning how to install it for free, you can save time and hassle and just pay them money to use software that installs this free toolkit for you.
Starting at U$57 for individuals, and U$147 for businesses, the purpose of the installer is to get people up and running as quickly as possible, while finding a concrete way to support the project financially. However if you’re patient or willing to learn, payment is not mandatory, and you can install this software on your own for free and use it as you see fit.
One of the advantages of getting into photogrammetry and ODM is that it enables a wide range of applications and means of using the data generated by the aerial images taken by a drone.
The software is also relatively easy to use, enabling all sorts of analysis or products generated from the initial set of photos.
The detail or depth involved depends upon how many photos you want to use. If numbered in the tens, then the detail is not that great, but the resources required to generate maps or models is not that onerous. Even in the hundreds, the average desktop computer should be able to have fun and generate decent composite imagery. However serious photogrammetric applications often involve thousands of images, and that’s where a dedicated workstation with loads of RAM is necessary.
In my own case I’m anticipating being in the tens if not low hundreds of photos, and am hoping that 16gb of ram is sufficient, which is the bare minimum that they recommend.
However my larger interest in the ODM toolkit is to explore what it is capable of. While this is true when it comes to the kinds of images or models I can create, it’s also true with regard to how the toolkit itself works and is being developed.
For example, NodeODM allows for models to be generated using a network of machines, and CloudODM allows for doing the same in the cloud (i.e. on someone else’s server infrastructure). This would allow for the use of thousands of images analyzed and integrated using more than just your own workstation. Knowing how to do this will be helpful as I scale up to analyzing larger forests and crop fields.
Self described as an open ecosystem, ODM runs the risk of having sprawling features and applications that compete for attention and resources. In addition to raising funds via a paid installer, the organization also recently raised money that helped users influence development priorities:
As maintainers of FOSS we receive hundreds of feature requests. Time is limited however, so how do you prioritize? Thumbs up on GitHub / forum polling is a “one-person one vote” system, which is not ideal. If you let a single organization sponsor the development of a feature, you get a “one-dollar one-vote” system, which gives too much power to the single organization.
Quadratic voting is a fairer system. After the funding period has ended, we know what the community values and we know what to prioritize.
Oh, and we raised some funds in the process, too! Win-win?
We’ve written a Future Tools issue in the past about quadratic voting, and it’s interesting to see the concept spread and be used as part of an open source software project with regard to their planned work.
This kind of user engagement helps to ensure the health of the project, and its relevance to the people who use it.
In this context, ODM is not the only photogrammetric application available, and the uses tend to be more professional than personal or hobby based. There are easier tools that can be bought, but do not have the same features or capabilities.
Which is why an open source toolkit like ODM can be so empowering. It creates a learning curve that offers a slippery slope for a novice user to become more experienced, even a professional, as they progress and advance their proficiency in using the tool.
This partly results from observing other people use the same tool for a wide range of professional purposes. For example in construction and architecture:
Another major application area is agriculture and forestry:
Geography in general is an obvious area in which this technology is used, but also a near infinite potential for art and creativity:
And of course there’s the application we referred to at the beginning of this issue, which is using imagery as a means of data collection, with maps being an efficient and potentially intuitive method of communicating or sharing that data.
I’m attracted to learning how to use ODM, not just to map out my own forest, but to learn how to help others do the same. This is where ODM starts to blend into the larger Open GIS world and the desire for people to connect with and use the GIS data that exists about their land and their community.
When we use maps we often don’t think about how they’re created, or the power they have in shaping our perception of the world. Yet if we can make it easier for people to make their own maps, maps that are created using their own tools, that will not only help shape people’s perception of their world, but also their agency, and their belief that they can do something about it. Or at the very least engage that world by trying to understand it.
What would you map if you could? What models would you make if you could?
Finally, check out this video on some creative applications for the ODM toolkit: