How Self-Learning Technology Could Disrupt Architectural Visualisation?
Technology continues to evolve and impress and as we endeavour to understand and keep up with the changes, it’s natural to have questions. An important one to consider is: How can self-learning technology disrupt architectural visualisation?
With the rise of self-learning software applications, through to artificially intelligent robotics, the issue of the impact it may have on the architectural visualisation aspect of the property development process is pertinent. It’s worthwhile remembering that it is also very difficult to just concentrate on only the elements of the 3D visualisation component that will be impacted because it is so interwoven throughout the overall process and the reality is that changes in the higher levels will have downstream consequences on each stage of 3D visualisation.
Because architectural visualisation, or 3D rendering, of unbuilt environments, is a combination of artistry and technical ability, there are some elements that could immediately be improved by automation and faster processing, thanks to intelligent software making decisions based on faster data access and greater body of knowledge, rather than relying on the experience of individuals or teams of 3D artists.
In the short-term future, one area where artificial intelligence (AI) could really help is in optimising processes where technical ability plays the greatest part. Imagine if building modelling was automated and fed into the computer from a hand-drawn architectural sketch and the system would then reference a database of optimal construction methods – and then deliver an architectural design that would already take costing and buildability of a development into consideration.
This could have major ramifications for the architectural industry, which would then flow on to our part – with the models we would receive being automatically optimised for rendering. In some cases, it could possibly replace the need for a render artist to produce detailed visualisations – something that would speed up the overall process.
The overall guiding assumption of this article is that humans, in the short to medium-term future, may still have greater abilities to innovate and be creative, whereas self-learning machines will be much better and make fewer mistakes in standardised processes, thus costing less and providing faster delivery than the human counterparts.
Impact of AI on Development Planning Visualisation
When it comes to marketing 3D visualisation images and interactive collateral, the level of technical information required is quite different to what is needed when visualising unbuilt spaces for development planning.
Creativity and art direction play a much greater role in off-the-plan marketing 3D renders versus planning 3D visuals. This is where self-learning technology could most assist and speed up the 3D visualisation delivery. Whilst the sophistication in software will continually improve, the greatest delay for us is, typically, in obtaining the initial site information and then converting it to a usable form for rendering.
Currently, when initial site information is collected, the main need is to coordinate a planning representative to confirm camera angle direction and a photographer to obtain the photo, then organise a site-surveyor to survey the location of the camera point. This often causes delays and creates opportunities for inaccuracy. But as the first steps of obtaining site information, such as site-survey and photography, are quite straightforward, in terms of their specific requirements, the future could be that these will be obtained by robots, which could be directed without even leaving the office.
Then, if the 3D modelling of the surveyed information could be available shortly after, as site-survey becomes more comprehensive and produced in 3D form with the photographs taken at the same time, this should speed up 3D visualisation delivery and, hopefully, improve the overall accuracy also.
Impact of AI on Off-the-Plan Marketing 3D Visualisation
Although there are a number of technical and process-driven aspects in the off-the-plan marketing space, the impact of creativity and innovation is a lot greater, and these are much harder to program and adapt to self-learning machines at this stage of our technological advancement.
Taking the image above as an example, the high-level objective of the 3D visualisation was to be able to communicate the overall layout of this complicated development and create an appealing presentation for the future residents. If self-learning machines were guiding the project, the decision-making process of interpretation of architectural plans – through to the selection of the camera angle and then to deliver an effective artistic representation – would be an extremely intricate process and unlikely to result in the solution achieved.
However, where technology can really help in the marketing solution is in the marketing floorplans delivery and production.
The process of converting marketing floorplans from AutoCAD or similar CAD packages into a graphic software, such as Illustrator, is much the same for every project. If an artificially intelligent machine could be fed the initial CAD files and deliver an accurate, yet graphically-presented, set of floorplans that were void of human error (as floorplans require an enormous amount of double-checking), the ability to have this available would be hugely beneficial.
Currently, the arduous process of cleaning up AutoCAD files takes an enormous amount of time and the process of ensuring accuracy for all small details across many floorplans leaves lots of room for errors to be made. On the other hand, computers are much better at delivering accurate deliverables – especially if the process is essentially the same – with the ability to isolate walls and other fixed elements and then convert them to an approved template and populate with symbolic scenery like plants and furniture pieces.
Artificial Intelligence in Custom 3D Visualisation Projects
In custom 3D visualisation, I forecast a lot of efficiencies can be gained by intelligent 3D library management and cataloguing of objects. What tends to happen most of the time for custom 3D visualisation, is that the clients want to leverage existing assets and bring their own products into existing scenes.
Over the 10 or so years of Eagle Vision, our collection of objects is quite considerable, however, its management has room for improvement. Furthermore, as new software versions are introduced, old versions of the models need updating and, often, they are rendered obsolete, as the time that it would take to update them is far longer than getting a new, similar object from another 3D library resource.
Automated and Intelligent Design through Artificial Intelligence
Although outside of the realm of architectural visualisation, self-learning applications could optimise building design. Ability to access a huge amount of data and process it rapidly could allow for considerably more agile development of planning applications and innovative solutions to current building constraints.
Big data analysis, such as taking into account personal needs of future residents based on demographic analysis, and, on the other hand, applying planning controls, is the ultimate dream for designers. However, in the immediate future, there is still a lot of fear around being able to control artificial intelligence robotics and software applications. Personally, I predict a 10-15 year timeframe before we see a lot more of this as common practice.
What is certain is that the future is very exciting and holds a lot of opportunities for the intersection of architecture and graphics for self-learning machines to speed up our work and help innovate virtual presentations to the next level.
To your development success,
LREA, BEng (Mech with Honours) / BTech (Industrial Design), VPELA