I discussed in a recent article the perception layer, the first element in the Geospatial 2.0 stack. Perception maps closely to our human senses: sight, hearing, smell, touch. From these raw inputs we begin to form insights (“I smell smoke coming from outside”). Those insights, in turn, guide intelligent decisions (“Maybe I should investigate in case my house is threatened”). That might lead us to action. This characterizes our 2.0 flow:
Context → Insight → Intelligence → Action (C-I-I-A)
Human cognition mirrors this flow: senses → understanding → reasoning → acting.
As the world around us changes, so too does our perception. But it struck me this week that perception is truly a data input. Our brains take that input and generate context and insight. This is the same in the 2.0 world. That recognition has made me, this week, refine my definition of the Perception Layer – raw signals from sensors, satellites, IoT, drones. So where is that data transformed into context and insight?
This is the Translation Layer (previously I had this as part of the Perception Layer). This is where we add semantic enrichment to the raw data. In practice, it’s the step where point clouds, imagery, and sensor feeds are transformed from ‘just data’ into structured meaning – what the objects are, how they relate, and why that matters.
We are talking here about knowledge graphs. For many readers, that may be a new term, but think of it like moving from a photo album to a family tree: the people are the same, but the tree shows who they are and how they connect. In the same way, knowledge graphs convert changing perception (data) into meaning (context and insight), which can then be passed on to the intelligence layer (reasoning) for decision-making and execution (action).
Open source tools like Neo4j can help generate these knowledge objects, making it easier to structure relationships directly from spatial data and keep them dynamic as the underlying data changes.
Let me add a little more here and provide a concrete example. A 3D point cloud is part of the human visualization layer. In the 2.0 world instead of a visual map, a 3D version of the Pantheon for example, we generate a knowledge map. Specifically, a point cloud to knowledge graph (the “glue” between perception and reasoning). In other words we add meaning to the 3D point cloud geometry – “these are columns,” “this is load-bearing,” or “this is eroded stone.” This ‘understanding map’ is then fed to a reasoning engine like VERSES Genius which then adds that new 2.0 dimension – intelligent decision-making: – Cultural Heritage: Monitor erosion over time, recommend preservation strategies. – Construction/Engineering: Compare as-designed vs. as-built models, flag deviations, auto-generate rework orders. – Defense/Logistics: Scan a port or facility, reason through vulnerabilities, simulate supply chain disruptions in real time. – Insurance: Feed 3D scans of damaged property into reasoning engines to automate claims assessments with explainable outputs.
This turns point clouds into living knowledge systems that can explain, adapt, and act.
In many ways, this is the real promise of Geospatial 2.0: a shift from data that simply describes the world to systems that understand it and can act on that understanding in real time. Perception gives us the raw signals, the Translation Layer turns those signals into meaning, reasoning engines generate intelligence, and agentic systems carry decisions into action. When these layers work together, we move beyond maps and dashboards to living, adaptive knowledge systems. And that is where the breakthroughs will come — not just from bigger datasets or faster models, but from the ability to continuously perceive, reason, and act as the world itself changes.