Photogrammetry is the digital and photographic capture of real-world objects and spaces in three dimensions. Photogrammetric captures use multiple digital images mapped spatially to create 3D models. There are established high-definition, professional workflows for creating photogrammetric scans, which I did familiarise myself with during the study. However, I was primarily interested in workflows that enabled rapid content creation using consumer-level software. I was working from the point of view of a content creator responding to events or news, a journalist. As such, my main focus during the study was on the use of mobile apps and imaging to create 3D models.
I embarked on a series of photogrammetric mini-projects. These were defined into three broad types:
1) photogrammetric captures of objects
2) photogrammetric captures of spaces (interior and exterior)
3) photogrammetric captures derived from drone footage of both spaces and objects.
I’m reflecting on this as one project because the skills and knowledge accumulated built on each of these in order.
At the beginning of the project, I was most interested in learning techniques and processes, capturing any object in the best possible resolution I could, given limited resources and time. I did not intend to replicate professional 3D development workflows but to arrive at an acceptable workflow for rapid content creation. As a result, the first objects I captured were incidentally available to hand. Later, as the project progressed, I created models and scans of newsworthy areas and spaces relevant to the study (for example, drone models of Dealey Plaza in Dallas).
Feelings
I encountered some diminishing returns with my approach. Initially, I achieved excellent results capturing smaller objects, which could be incorporated into 3D models and spaces. As the project expanded, I tried to capture more intricate interiors and exteriors, only to quickly realise that photogrammetry is most effective when there are clear boundaries and the spaces captured are open and uncluttered. The more complex the interior, the more challenging it is to obtain a satisfactory capture. Exterior spaces are even more difficult to capture with a mobile phone. Some iOS devices have built-in lidar, a bonus for photogrammetric capture of interiors and rooms. Nevertheless, even in those cases, results vary. Fortunately, the limitations of the physical space available for users to navigate in VR mean it is better to capture simpler or smaller spaces rather than larger, more complex areas.
I achieved satisfactory results using drone footage, especially when photographically capturing landmarks and notable buildings. I also got some excellent models of exterior spaces this way. Some of the outcomes from those projects were gratifying and potentially valuable for further study in the rapid development of spatial storytelling in immersive journalism.
I focused almost entirely on mobile capture, primarily using PolyCam and some other mobile tools that perform a similar job. However, none of them matched Polycam’s standards. I also utilised Reality Capture, which, combined with high-definition photography, produces far superior results to some of the PolyCam scans I created. To adhere to a budget and consider immersive journalism as an everyday practice, I imposed these limitations on myself, consequently impacting my learning. I aim to rectify these limitations in the future.
Evaluation
I have learned to work within the limitations of the tools. For example, I have developed a workflow where I default to using the room capture feature in PolyCam to sketch out a 3D model of an interior space and then import that into SketchUp for editing. Polycam’s room feature uses LiDAR to measure and scan a space and intelligently fills in that space using primitive shapes. In a 3D architectural design tool like SketchUp, we can place objects, apply textures, and improve the quality of the model.
I believe spaces with clear boundaries work best for outdoor photogrammetric capture. For example, I have effectively captured good models of enclosed gardens and alleyways. My approach will be to compartmentalise these individual spaces as objects.
A limitation of using mobile apps like Polycam for photogrammetric processing is that the app does much of the work, whereas a professional-quality workflow using Reality Capture requires more intervention from the designer or developer. This leads to a greater and deeper understanding of the process, the ability to identify what elements are and are not a result of user error, and how we improve our scans.
Application
The project demonstrates that it is conceivable to rapidly develop immersive spaces using photogrammetry, which can be used for storytelling. Photogrammetry can augment both the creation of spaces and objects. However, it’s also clear that in most cases, for rapid development, a mixture of techniques might be best deployed. This may include the combination of 3D modelling of spaces, texture mapping using situational photography, and photographic capture of objects. More work will need to be done as part of a postdoctoral project.
Conclusions
I’ve been satisfied overall with the work I’ve done so far—satisfied enough, at least, to make this the focus of practical projects going forward. These will explore or continue to explore methods for creating authentic spaces for immersive journalism that can be quickly assembled as the foundational base for storytelling using a combination of photogrammetry, photography, video, and 3D modelling.