Explore

VIZZIO.AI

Ultra Large-Scale 3D RECONSTRUCTION of DIGITAL CITY MODELS of the WORLD BUILT BY ALGORITHMS & A.I.

Last edited 6 minutes ago by CEO Vizzio

VIZZIO maps and models WORLD cities into high fidelity photorealistic 3D models, backed by an AI automated pipeline that segments and understands the semantics of satellite & street images, videos, point clouds and generates DIMENSIONALLY ACCURATE geo-specific & geo-typical parametric and hyper-photorealistic 3D digital twins. Complete Globe Coverage. No Aerial Photogrammetry needed.

Our blazing fast AI-powered 3D City Modelling workflow has a canny understanding of what different buildings look like from above, even in suboptimal lighting and incomplete imagery.

VIZZIO's 3D Digital Earth Modeling system utilizes multi-petabytes of satellite imagery and geospatial data, in near real-time, on a planetary scale, employing deep learning to create a semantic, machine-readable, and photorealistic 3D digital representation of the Earth.

⁠

⁠

VIZZIO builds the world’s leading 3D modelling and visualisation software, enabling city-scale digital twins and mission-critical dashboards for governments and enterprises. We are a global leader in the production of highly detailed digital replicas of urban landscapes, enriched with semantic 3D data, made readily available for an array of uses including gaming, simulations, training, city planning, defence, sustainability and much more,. No drones or aerial photogrammetry needed.

BEING ABLE TO CAPTURE ANYWHERE GLOBALLY with HIGH ACCURACY IS A MASSIVE GAME CHANGER

VIZZIO’s patented AI technology revolutionises 3D city modeling, creating highly accurate digital twins of any location on Earth using satellite imagery and data from leading geospatial providers worldwide. This innovative process ensures geometrically precise and positionally accurate models that are always up-to-date, delivered through an incredibly fast, AI-driven workflow. No drones, aerial photogrammetry, or government permits required. Transform urban mapping with unparalleled efficiency and detail.

⁠

Our faithful building facade texturing process combines artificial intelligence, heuristics, and procedural methodologies. We used a UNET neural network, trained on 1500+ different cities across the globe, to classify each pixel into other classes, such as building, shadow, and eight different roof shapes. This unique and patented technique creates a photorealistic, life-like environment from satellite imagery and exogenous data. We can also integrate other relevant data, such as GIS, street names, points of interest, and zoning information. This integration enhances the accuracy and utility of the model.

⁠

VIZZIO Planetary-Scale 3D City Mapping. Powered by Satellite Imagery & AI. MASSIVE 82 3D cities modelled - 3.5 PB DATA - 953,293 square kilometers covered. Fusing multi-resolution and multi-sensor Earth Observation & Satellite data with ML and AI to provide near-real time actionable geospatial insights and city-scale 3D modelling.

⁠

Singapore 3D Screen Saver (4K) - Aerial view to Street view to indoor

⁠

Our planet-scale mapping and visualisation also work across any devices

THE COMPANY

⁠

We anticipate doubling our growth in 2024, driven by several key factors. Significant contributions are expected from the sales of POLYTRON.AI cameras, which have shown strong market demand. Additionally, new product launches will serve as major revenue drivers, further boosting our financial performance. We also foresee expanded contracts and business opportunities in the Middle East, APAC, and Europe, which will play a crucial role in achieving this growth.

WHAT WE DO

We bring multi-modal AI-POWERED 3D modelling and visualisations to consumers and businesses.

We scan OBJECTS

We map SPACES (indoor/outdoor)

We model CITIES

⁠

THE OPPORTUNITY

Recent advancements in Remote Sensing and Photogrammetry have emphasised the significance of modeling Earth's surface and extracting topographic information. Over the past 15 years, high-resolution satellites, offering images with spatial resolutions of 1 meter or finer and stereoscopic coverage, have emerged as viable alternatives to traditional aerial photography for urban mapping and detailed 3D building modeling. This shift is primarily due to the prohibitive costs associated with conventional aerial photography and drone-photogrammetry, which demand extensive manual labor for drone operation and significant computing resources to convert imagery data into 3D models.

Unlike Google and other Aerial Photogrammetry companies, we build DIMENSIONALLY-ACCURATE, PHOTOREALISTIC, REAL-TIME, SEMANTIC 3D DIGITAL TWIN OF CITIES via deep learning and satellite imagery on a global scale, backed by an AI automated pipeline that segments and understands the semantics of point cloud and generates usable parametric and 3D models.

⁠

⁠

The challenge of utilising UAV photogrammetry to construct 3D city models lies in the costly and time-consuming nature of the process. Furthermore, obtaining government permits for no-fly zones adds another layer of complexity. This costly and protracted procedure complicates the task of regularly updating the 3D map to ensure its relevance.

⁠

SLOW. EXPENSIVE and LABORIOUS Process of 3D mapping using UAV photogrammetry vs VIZZIO’s patent-protected city mapping methods using 2D satellite imagery & Deep Learning.

SLOW. EXPENSIVE and LABORIOUS Process of 3D mapping using UAV photogrammetry vs VIZZIO’s patent-protected city mapping methods using 2D satellite imagery & Deep Learning.

⁠

As of 2023, it is believed that only 5% of the planet has been mapped using LiDAR systems. Although LiDAR maps are incredibly useful, collecting the data is time-consuming and expensive. Building a LiDAR map of one region would cost millions of dollars and take several years to complete.

LiDAR mapping is also extremely resource-intensive. It requires pilots, expensive equipment, advanced software, and skilled technicians to capture the information and turn the point cloud data into something useful. What is more, LiDAR can quickly go out of date. If land use or land cover changes, then the information from the maps quickly becomes irrelevant.

But it is not the only way of creating 3D models. At VIZZIO, we have developed an accurate and efficient methodology for automatically extracting footprints of buildings and vegetation based purely on satellite imagery, then using shadows to calculate height and depth of objects on the surface. You can use this information to rapidly build 3D models in a very short space of time without needing any additional equipment. Our solution provides the three dimensions provided by LiDAR maps, but at a far lower cost.

⁠

⁠

VIZZIO's cutting-edge, patented technologies for city-scale satellite 3D mapping represent a significant advancement in geospatial intelligence. These innovative solutions allow for the creation of highly detailed and accurate 3D maps of entire cities, leveraging satellite imagery and advanced algorithms. This technology is crucial for applications in urban planning, disaster management, real estate, and navigation, providing unparalleled precision and insights into urban environments. VIZZIO's pioneering efforts in this field highlight our dedication to pushing the boundaries of what's possible in 3D mapping and spatial analysis.

⁠

Our groundbreaking AI technology is the driving force behind the creation of our hyper-photorealistic virtual world. Simply select any location on the global map that you wish to model, purchase the corresponding satellite images, and watch as our AI works its magic. The AI seamlessly transforms these images into a detailed and accurate digital replica of the chosen area, automating the entire process and delivering stunningly realistic results without the need for manual intervention. This patented approach allows for effortless and precise 3D world-building, revolutionising how we create and interact with digital environments, without the need for remote sensing method such as aerial photogrammetry.

We are creating a 1:1 scale virtual world that is an exact photorealistic replica of Earth, setting a new standard in digital city modeling. Unlike traditional methods that rely on procedural mapping or extensive human effort to model buildings and roads, our approach leverages advanced AI and satellite data. Our AI is capable of accurately interpreting every building, road, and structure to generate precise digital assets, eliminating the need for manual modeling. This results in a highly detailed and realistic virtual environment that mirrors the real world with unmatched accuracy and efficiency.

CROSS BROWSER

⁠

OUR SOLUTION

Our AI-driven, patent-protected large-scale 3D Reconstruction of buildings from stereo satellite images pipeline uses Deep Convolutional Network to extract contour polygons of buildings, and the combination of optimisation & computational geometry techniques to compute the Digital Terrain Model, a digital height model and to correctly estimate the position of the footprints. shadow & atmospheric cloud removal.

⁠

And we have overcome a number of problems including:

influence of atmospheric and solar illumination: shadow, shade, and haze

perspective viewing problems: occlusions and relief distortions

similarity between the building region characteristics and their background, including spectral, textural, and geometric properties, such as shape

⁠

Satellites provide us with a unique overview of our planet. About 5000 satellites are orbiting above us and about 700 of these specifically record earth observation data. As satellites continuously orbit the earth, they enable us to track physical trends, changes, and developments over long periods of time – and in great details. These satellites imagery also enable us to build photorealistic 3D digital twins comprising of digital surface model (DSM), digital terrain model (DTM), and digital elevation model (DEM), and document all the change and dynamics on the surface – and this can be used for a wide range of applications.

⁠

We model entire city in high fidelity photorealistic 3D via VHR sub-30 cm satellite imagery. Powered by Deep Learning AI.

We model entire city in high fidelity photorealistic 3D via VHR sub-30 cm satellite imagery. Powered by Deep Learning AI.

⁠

There are various types of earth observation remote sensing satellites we used for our geometric computation, each designed for specific purposes and equipped with different sensors. Imagery data are acquired from a constellation of low-earth orbit sun-synchronous satellites.

SAR

Optical Satellites

Thermal Infrared Satellites

Hyperspectral Satellites

Spatial resolution is defined as the size of the smallest object that can be reliably detected in an image. It is typically expressed in meters or centimeters per pixel. High spatial resolution means that each pixel represents a smaller area on the ground, resulting in more detailed images.

⁠

High spatial resolution allows for the detailed mapping of urban environments, capturing fine details such as individual buildings, roads, and vegetation.

Essential for applications requiring precise measurements, such as urban planning, infrastructure development, and environmental monitoring.

High-resolution imagery provides city planners with detailed views of urban landscapes, aiding in the design and development of infrastructure.

Enables the creation of accurate 3D models for visualisation and simulation purposes.

Detects structural details and changes in buildings and infrastructure, helping in maintenance and management.

Supports emergency response by providing up-to-date information on urban layouts and potential hazard zones.

Monitors green spaces, water bodies, and other environmental features within urban areas.

Assesses the impact of urban development on natural resources and ecosystems.

High spatial resolution is crucial for monitoring crop health, detecting diseases, and managing forests.

Allows for precise mapping of agricultural fields and forested areas.

Provides detailed imagery for assessing damage from natural disasters such as earthquakes, floods, and hurricanes.

Supports rescue and recovery operations by identifying affected areas and infrastructure damage.

Tracks changes in land use, deforestation, desertification, and other environmental phenomena.

Helps in the study and mitigation of climate change impacts.

The creation of dimensionally accurate, high-fidelity, fully textured 3D city models from satellite imagery and Synthetic Aperture Radar (SAR) data has become increasingly feasible with advancements in remote sensing and artificial intelligence (AI). These models are crucial for urban planning, disaster management, and various other simulation & visualisation applications. This document details the process of using Very High Resolution (VHR) satellite imagery and SAR data to produce such models, emphasising the importance of spatial resolution in achieving geometric accuracy and texture resolution.

⁠

Planetary-scale city modelling platform converts 2D satellite imagery into a semantic, high-performant, photorealistic, synthetic 3D digital twin of the entire planet. Powered by Satellite Imagery & Deep Learning

Planetary-scale city modelling platform converts 2D satellite imagery into a semantic, high-performant, photorealistic, synthetic 3D digital twin of the entire planet. Powered by Satellite Imagery & Deep Learning

⁠

To gather data for creating 3D models of cities, you have the option to use airborne, terrestrial, or space photogrammetry. These methods involve using aerial, terrestrial, or satellite images as the primary source of data to extract 3D point clouds from pairs of stereo images.

Super High-resolution 15 cm satellite images capture very detailed information of the terrain. This enhanced super resolution satellite imagery is obtained by applying proprietary AI Super-Scaler Imaging technology to native 30 cm resolution data. This technology intelligently increases the number of pixels through a complex mathematical model, resulting in a superior visual experience without changing the original ground sampling distance (GSD). And we are now employing this 15 cm HD satellite images to our 3D city reconstruction process, as demonstrated below.

⁠

Singapore 3D - 15 cm VHR Satellite Imagery 3D City Map Generated

The enhanced precision of our satellite images significantly improves our ability to detect buildings and calculate building envelopes, as well as generate Digital Surface Models (DSM), Digital Elevation Models (DEM), and Digital Terrain Models (DTM). The use of 15 cm high-definition (HD) imagery provides extra detail, which greatly benefits automated rooftop analysis and the extraction of terrain features. This higher resolution allows for more accurate and detailed assessments of both man-made structures and natural landforms, facilitating more precise urban planning, environmental monitoring, and infrastructure development. The additional granularity helps in identifying finer elements on rooftops and landscapes that were previously not discernible with coarser imagery, enhancing the overall quality and usefulness of the spatial data for 3D City Modelling and ReConstruction.

COMPETITIVE ADVANTAGE

Our AI-powered 3D Modeling & Visualisation Technologies stand at the forefront of the industry, setting new benchmarks for efficiency, innovation, and affordability. Engineered for speed, our software leverages cutting-edge technology to ensure lightning-fast processing of 3D modeling tasks. This significant reduction in workload and enhancement in productivity deliver superior performance and usability. All these advantages come at a fraction of the cost of our competitors, making it the smarter choice for businesses looking to gain a competitive edge.

⁠

⁠

We have also secured a WORLDWIDE patent for a novel urban mapping technology, which represents a significant innovation in our approach to urban cartography. This technology integrates a comprehensive workflow, enabling the detailed mapping of every city on Earth without relying on the deployment of drones or aircraft for data acquisition in specific locales.

⁠

McKinsey's internal report comparing the various aerial mapping method against VIZZIO AI-Powered satellite-based 3D city reconstruction approach

McKinsey's internal report comparing the various aerial mapping method against VIZZIO AI-Powered satellite-based 3D city reconstruction approach

⁠

The advancement in satellite imagery resolution has significantly improved the feasibility and quality of 3D mapping from space. VHR satellites can capture imagery with a resolution of less than 0.15 meter per pixel, allowing for the highly-detailed mapping of urban landscapes.

Procedural 3D City Mapping vs Satellite-based 3D Reconstruction

⁠

On the LEFT, you'll see the conventional method most people use to create 3D city models, which often involves labor-intensive processes and extensive human effort.

On the RIGHT, we showcase our revolutionary Auto-Generated 3D city reconstruction based on satellite imagery & Deep Learning. Our technology seamlessly transitions from aerial views to street-level perspectives and even indoor environments. This process turns these photorealistic city models into live, interactive experiences, creating a hyper-realistic replica of our planet. Our approach brings every detail of both outdoor and indoor spaces to life, offering an immersive and dynamic digital twin of the Earth.

ACCOLADES

VIZZIO is ranked #1 Most promising startup in Asia by The Information (Silicon Valley’s Most Influential Tech Media), in it's annual

#TheInformation50⁠

most promising startups globally.

⁠

⁠

PATENTS

VIZZIO.AI has made significant strides in intellectual property protection, having filed numerous patents with the United States Patent and Trademark Office (USPTO). We are on track to expand our portfolio to a total of 50 patents by the end of 2024. This growth in our patent holdings underscores our commitment to innovation and solidifies our position as a leader in the AI and technology sectors.

⁠

VIZZIO.AI's filed patents USPTO, growing to 50 patents by end 2024.

VIZZIO.AI's filed patents USPTO, growing to 50 patents by end 2024.

⁠

WORLD's FIRST Patents Awarded on GeoSpatial 3D Mapping of Cities using Satellite Images. Powered by AI

We are excited to share that on 27 February 2024, VIZZIO Technologies Pte Ltd has received a groundbreaking USPTO patent for our innovative approach in integrating 2D satellite imagery with SOTA AI technology to produce geometrically correct and highly detailed 3D city models. This milestone is a significant achievement for VIZZIO, as it not only highlights our pioneering efforts but also paves the way for the exploration of these advanced 3D mapping technologies across various applications such as smart city initiatives, enhancing urban safety, and promoting circular city development.

⁠

WORLDWIDE FIRST PATENT on 2D Satellite Mapping to 3D City Model

WORLDWIDE FIRST PATENT on 2D Satellite Mapping to 3D City Model

⁠

This is our most significant patent filing to date. It encompasses 15 methods and claims detailing our process of transforming 2D satellite images into 3D city models. Notably, it represents a world-first by encapsulating our entire workflow for mapping every city on Earth without the necessity of deploying drones or planes to specific locations. This technology stands as a pivotal component within our AI modelling pipeline.

This patent isn't just a recognition of our past accomplishments; it's a springboard for future innovation and collaboration. As we continue to explore new frontiers and tackle complex challenges, we are confident that our patented technology will pave the way for transformative advancements in urban planning, environmental sustainability, defence, simulation and beyond.

Together, we have demonstrated what can be achieved when vision, technology, and teamwork converge to push the boundaries of what's possible.

⁠

ABOUT US

⁠

⁠

VIZZIO develops leading-edge 3D modeling and visualisation software, specialising in the creation of city-scale digital twins and mission-critical dashboards for governments and enterprises. This global leader in digital replication uses semantic 3D data to craft highly detailed replicas of urban environments for diverse applications such as gaming, simulations, training, city planning, defense, and sustainability, without the need for drones or aerial photogrammetry.

We are a venture-backed startup creating a radically new kind of 3D modelling and visualisation technology. Powered by Deep Learning. We constantly experience things spatially. Our brains are constantly calculating distances, sizes, and volumes to understand the world around us. Videos fall flat and aren't interactive. Instead, an interactive, immersive 3D content gives more than visuals.

We are recognized globally for producing meticulously detailed, precisely measured 3D digital building models using advanced satellite imagery and artificial intelligence — eliminating the need for drones or aerial photogrammetry.

Our ambition is to forge photorealistic digital replicas of urban landscapes, enriched with semantic 3D data, made readily available for an array of uses including gaming, simulations, training, city planning, defence, sustainability and much more. Utilising advanced machine learning techniques, our algorithms deftly parse raw satellite imagery, while our geographic pipeline extracts semantic information about what's in the world. With each digital twin we build, our AI process becomes increasingly precise and reliable, surpassing the capabilities of aerial photogrammetry, producing positionally and geometrically accurate 3D maps.

We are creating the authoritative whole-Earth digital twin, built by AI, one city at a time. Real-world cities as immersive digital twins and rich 3D data, instantly accessible for games, simulations, training, architecture, and beyond. Our machine learning algorithms understand raw imagery data from satellites while our geographic pipeline extracts semantic information about what's in the world. The more data we have, the more accurate our system becomes.

Utilising satellite imagery and data from top geospatial providers, we can accurately model any location worldwide. The models are geometrically and positionally precise, updated through a fast, AI-driven process. The texturing of building facades is achieved using a neural network trained on imagery from over 1500 cities, classifying details like building elements and roof shapes to create photorealistic environments. Additional data integration, such as GIS and zoning information, further enhances the model's accuracy and utility.

Advanced machine learning and a geographic pipeline allow VIZZIO to efficiently parse raw satellite imagery and extract detailed semantic information, continually refining the accuracy of their digital twins and outperforming traditional aerial photogrammetry methods.

The possibilities for 3D mapping are limitless — ranging from wireless planning, city planning, infrastructure management, disaster management, and environmental monitoring to applications in public health and academic research. VIZZIO 3D city maps are defined by superior visualisation performance, cost-effectiveness, short delivery times, and unmatched flexibility. The quality of this 3D data will only improve as the number of available satellites increases and satellite-mounted sensors become more sophisticated.

As an industry, Earth Observation has made great progress in understanding our planet in 2D. VIZZIO’s proven city-scale 3D mapping of the world will open up the market for space-based EARTH Observation, unlocking billions of dollars of demand for timely, up-to-date 3D city models.

⁠

We are a team of creative technologists on a mission to reimagine the art and science of 3D.

We are a team of creative technologists on a mission to reimagine the art and science of 3D.

⁠

PROJECTS

Customer Project Page

Link⁠

⁠

FOXCONN INDUSTRIAL INTERNET and VIZZIO redefine Smart Manufacturing with Global Digital Twin Collaboration for the World’s First AI Server WEF Lighthouse Factory

Link⁠

⁠

NATIONAL UNIVERSITY SINGAPORE SMART CAMPUS & IOC Digital Twin

Link⁠

⁠

Economic Cities & Special Zones Authority (ECZA), Saudi Arabia - Won National Innovation Award 2023 Awarded by Digital Government Authority & Nomination to 2 United Nation Award representing Saudi Arabia

Link⁠

⁠