Evolution of Reality Capture

Chapter 1 · Introduction - What is Reality Capture?

Definition and scope

Reality Capture is simply the set of technologies that create a faithful and measurable digital copy of the physical world. The term covers a very broad spectrum: from the terrestrial laser scanner set on its tripod in a factory to drone photogrammetry over a construction site, through the LiDAR sensors embedded in our smartphones, 360-degree cameras and mobile mapping systems. The fundamental objective never changes: transforming the real world into exploitable, measurable, shareable and analysable digital data.

What distinguishes Reality Capture from traditional 3D modelling is that it starts from reality. A 3D modeller manually builds geometries on screen. Reality Capture captures what already exists. The data produced come in various forms: point clouds (these collections of millions, sometimes billions of XYZ coordinates), textured 3D meshes, georeferenced orthophotos, navigable 360-degree panoramas, or as-built BIM models. Together, these data form the foundation of what the industry now calls the "digital twin".

Why this revolution matters

In twenty-five years, reality capture has gone from a technological niche reserved for surveyors, industrialists and archaeologists to a strategic tool used in architecture, engineering, construction, industry, infrastructure, heritage, insurance and real estate. A cautious reference is the global 3D scanning market, estimated at $4.28 billion in 2024 and projected at $7.51 billion in 2030, with a CAGR of 10.1% for 2025-2030 according to Grand View Research.

How to explain this acceleration? Several factors converged at the right time. First, the spectacular fall in hardware costs: a laser scanner cost EUR 200,000 in 2000, it costs EUR 20,000 in 2025, and a smartphone LiDAR is integrated into a device at EUR 1,000. Then, the explosion of computing power via cloud and GPUs. The miniaturisation of sensors, too, which allowed a scanner to fit in a backpack, then in a pocket. The democratisation of drones, obviously. And more recently, the integration of artificial intelligence into processing chains.

Chapter 2 · Before 2000 - Prehistory: the roots of 3D scanning

Military and space origins

The history of Reality Capture begins well before the year 2000. LiDAR (Light Detection and Ranging) was invented in the 1960s, shortly after the development of the laser. The first applications were military and space-related. In 1971, during the Apollo 15 mission, a laser altimeter aboard the command module mapped the lunar surface, producing the first extraterrestrial LiDAR data in history. This pioneering experiment already demonstrated the potential of laser measurement for large-scale mapping.

Anecdote. The word "LiDAR" is a portmanteau of "light" and "radar". For decades, purists debated the correct spelling: LIDAR, Lidar or LiDAR? The American Meteorological Society ruled in favour of "lidar" all lowercase, as a common noun. But the Reality Capture industry has predominantly adopted "LiDAR" with mixed capitalisation. The debate still raises smiles at conferences.

Photogrammetry: a century of history

Photogrammetry is much older than LiDAR. The term was invented in 1867 by the German architect Albrecht Meydenbauer, and the genesis is delightful. Meydenbauer had nearly fallen from the cathedral of Wetzlar while attempting to take manual measurements of the facades. This narrowly avoided fall drove him to imagine a method for measuring buildings from photographs, without risking his life. It is often cited as the "founding moment" of the discipline: a science born from the need to no longer risk one's life to measure a building.

In the twentieth century, aerial photogrammetry became the standard tool for national mapping. Aircraft equipped with metric cameras systematically photographed the territory, and specialised operators produced topographic maps using stereocomparators, those optical instruments that allow perceiving relief from image pairs. This process, entirely analogue and manual, required years of work and considerable expertise.

The pioneers of terrestrial 3D scanning

The first terrestrial laser scanners appeared in the 1990s. In 1994, Cyra Technologies was founded in Oakland, California, by Ben Kacyra, an American civil engineer of Iraqi origin. Kacyra's story is a beautiful one. He had been profoundly affected by the destruction of cultural heritage in Mosul during the Gulf War. His primary motivation was not industrial but cultural: he wanted to create a tool capable of digitising and preserving threatened monuments.

Cyra's first product, the Cyrax 2400, launched in 1998, was a terrestrial laser scanner capable of measuring thousands of points per second with centimetre-level accuracy. Despite its weight (over 25 kg), its price (approximately $200,000) and the complexity of its use, it represented a considerable advance. For the first time, one could capture the complete geometry of a space in a few hours rather than in days or weeks of manual surveying.

The story of Ben Kacyra. After making his fortune with Cyra Technologies (acquired by Leica Geosystems around 2000-2001), Kacyra founded CyArk, a non-profit organisation dedicated to the 3D digitisation of threatened world heritage. CyArk has digitised hundreds of iconic sites, from Pompeii to Angkor Wat, including Mount Rushmore and the Great Mosque of Djenne in Mali. His 2011 TED Talk, "Ancient wonders captured in 3D", helped raise public awareness of the potential of 3D scanning for cultural preservation. A man driven by a mission, not just by a market.

The state of the digital world in 1999

To measure the distance travelled, one must recall the state of the digital world at the end of the twentieth century. It is a dizzying exercise. In 1999, the average internet speed in France was 56 kbit/s (a dial-up modem making that distinctive noise). Google was one year old. Wi-Fi had just been standardised. The first iPhone would not arrive for another eight years. Cloud computing simply did not exist: Amazon Web Services would not launch until 2006. Transferring a 500 MB point cloud via the internet would have taken over 20 hours. The idea of a real-time 3D cloud visualisation platform was science fiction.