dark
Trending Tags
A serene view of Alcatraz Island, featuring its historic buildings and a lighthouse, surrounded by calm waters reflecting the cloudy sky during twilight.

RealityScan 2.0: Epic Gives RealityCapture a New Name and New Tricks

17.06.2025
Epic overhauls RealityCapture into RealityScan 2.0, complete with AI masking, LiDAR fusion, and a mobile-first workflow focus.

Epic Games has launched RealityScan 2.0, a comprehensive rebranding and feature update for its formerly desktop-centric photogrammetry solution, RealityCapture. The rebrand comes with a refreshed focus: mobile-first workflows, automated AI tools, and integrated support for aerial LiDAR. The move repositions the app to better serve users across industries from VFX and game development to geospatial analysis and AR/VR content creation.

From Capture to Scan: A Unified Ecosystem

The new branding consolidates Epic’s photogrammetry tools under the RealityScan name: desktop software becomes RealityScan 2.0, while the mobile companion app is now called RealityScan Mobile. Both are positioned as part of Epic’s ecosystem supporting Unreal Engine workflows.

While the underlying reconstruction engine remains based on RealityCapture technology, the update improves usability and broadens appeal beyond photogrammetry veterans. Epic’s aim: turn scanning into a faster, simpler step in asset pipelines—without sacrificing accuracy.

A close-up view of a red apple displayed on a black surface, with multiple images of the apple in various angles arranged in a grid. A progress message showing 'Generating Masks' is visible in the center of the image.
A close-up view of a red apple displayed on a black surface, with multiple images of the apple in various angles arranged in a grid. A progress message showing 'Generating Masks' is visible in the center of the image.
A detailed 3D digital model of an apple surrounded by small reference images in a software interface. Two actual apples, one with a red hue and one with a pinkish tint, are displayed on either side of the model against a black background.
A detailed 3D digital model of an apple surrounded by small reference images in a software interface. Two actual apples, one with a red hue and one with a pinkish tint, are displayed on either side of the model against a black background.

AI Masking: Segmentation Without the Pain

RealityScan 2.0 introduces an AI-powered segmentation tool that automatically isolates foreground objects from cluttered or textured backgrounds. A scan of fruit on a tabletop, for instance, no longer needs manual cropping or custom masking—machine learning handles the dirty work.

This AI model performs per-image semantic segmentation, making it particularly useful when scanning single assets like props or small objects. It’s a speed boost and a time-saver, especially for teams who need to generate quick scans for previz, virtual production, or rapid prototyping.

https://cms-assets.unrealengine.com/cm6trzq52049r07ohhff9djhz/resize=width:1200/output=format:webp/cmbr7ml5s3mko07oe8dlyejt6

Better Alignment = Fewer Headaches

Another under-the-hood improvement: enhanced image alignment algorithms. Especially for objects with low-feature surfaces—like smooth metals or organic shapes—the updated alignment makes it easier to get stable reconstructions without constant tweaking of parameters. This shift reduces the failure rate of initial reconstructions, offering more consistent results out of the box.

A glowing apple shape made of tiny, colorful particles against a black background, with text below reading 'NEW FEATURE SMARTEST ALIGNMENT'.
A glowing apple shape made of tiny, colorful particles against a black background, with text below reading 'NEW FEATURE SMARTEST ALIGNMENT'.
A comparison of two 3D models of an apple. The left model is the old version, featuring a textured, incomplete surface, while the right model shows a smoother and more polished apple marked as "RS 2.0 Alignment."
A comparison of two 3D models of an apple. The left model is the old version, featuring a textured, incomplete surface, while the right model shows a smoother and more polished apple marked as "RS 2.0 Alignment."

LiDAR Joins the Party

Perhaps the biggest technical expansion in RealityScan 2.0 is LiDAR data integration. The software now supports combining aerial LiDAR scans with terrestrial or aerial photogrammetry—a feature previously limited to the desktop version. This makes RealityScan 2.0 applicable to a broader range of production types, including environmental reconstruction, terrain modelling, orthophoto generation, and site mapping.

A 3D model of a rocky island showcasing a complex structure, with roads and greenery visible, under dramatic lighting against a dark background. The scene captures a mixture of natural and man-made elements.
A 3D model of a rocky island showcasing a complex structure, with roads and greenery visible, under dramatic lighting against a dark background. The scene captures a mixture of natural and man-made elements.

During Geo Week, Epic showcased a full environment scan of Alcatraz combining aerial images and LiDAR. While niche, the workflow will appeal to studios and teams creating virtual environments or working on large-scale simulations.

A 3D model of an island displayed on a computer interface, showing detailed terrain features, water surrounding the land, and graphical data points scattered around. The viewing area includes a grid for spatial reference.
A 3D model of an island displayed on a computer interface, showing detailed terrain features, water surrounding the land, and graphical data points scattered around. The viewing area includes a grid for spatial reference.
Aerial view of a fortress with multiple buildings and pathways, surrounded by greenery and water, displayed on a dark interactive interface with navigation tools.
Aerial view of a fortress with multiple buildings and pathways, surrounded by greenery and water, displayed on a dark interactive interface with navigation tools.
An aerial 3D rendering of a small island featuring a building complex surrounded by lush greenery and rocky shores. The digital environment shows a grid background, indicating a design software interface. Blue water gently surrounds the island.
An aerial 3D rendering of a small island featuring a building complex surrounded by lush greenery and rocky shores. The digital environment shows a grid background, indicating a design software interface. Blue water gently surrounds the island.
https://cms-assets.unrealengine.com/cm6trzq52049r07ohhff9djhz/resize=width:1200/output=format:webp/cmbr7n2cu3ngm07oehsm5b0dy

Heatmaps for Mesh Confidence

RealityScan 2.0 now includes visual quality inspection tools in the form of data completeness heatmaps. These maps display sampling coverage over scanned objects, helping users identify weakly covered areas before proceeding to meshing or texturing.

A computer screen displaying a 3D model of a sphere with a gradient color scheme. The overlay text reads, 'Coloring Vertices,' with a completion status indicating 'finished in 1 second' in the center. The background shows a grid pattern, typical of 3D design software.
A computer screen displaying a 3D model of a sphere with a gradient color scheme. The overlay text reads, 'Coloring Vertices,' with a completion status indicating 'finished in 1 second' in the center. The background shows a grid pattern, typical of 3D design software.
A 3D visualization of a spherical object displaying a gradient color map, with segments in green, yellow, and red. The object is rendered against a dark grid background, alongside a user interface showing analysis tools and data metrics.
A 3D visualization of a spherical object displaying a gradient color map, with segments in green, yellow, and red. The object is rendered against a dark grid background, alongside a user interface showing analysis tools and data metrics.

This feature isn’t just cosmetic. It can prevent wasted processing time and unexpected reconstruction errors, particularly when dealing with intricate or occluded geometry. For asset teams in production, that’s a tangible win.

Real Artists, Real Feedback

Initial community responses, particularly from photogrammetry and AR developers on Reddit, suggest cautious optimism. Users welcomed the improvements in masking and alignment, though many voiced the usual caveat: these tools need to be tested before being trusted.

A Tool That Still Needs Vetting

While the update introduces smart automation and expanded formats, the core of photogrammetry hasn’t changed. Users will still need to validate mesh fidelity, texture resolution, UV cleanliness, and topology before production integration. As with all tools promising simplified workflows, RealityScan 2.0 needs a test drive in real-world conditions.

For professionals in VFX, game development, digital twins, and AR/VR, the draw is clear: faster, lighter, and more automated photogrammetry. But as always, performance matters more than promises.

Learn more here: RealityScan

Bela Beier Author