Scape.io — Changing The Way We See The World

Introduction

Scape Technologies was founded in 2016 by CEO Edward Miller, who has a strong background in interactive imagery, and CTO Huub Heijnen, who was a researcher in robotics and machine learning. The Scape team are using cutting edge AI to create large scale 3D maps of real cities, enabling camera devices, like your phone, to recognise their surroundings and location far more precisely than GPS alone. At the beginning of 2019, Scape had raised $8m in seed funding and are now providing their mapping service to select developers in London.

Building a city

But how does it work? How would your phone camera be able to recognise any building or street in London just by looking at it? Well first, the Scape team had to build a 3D map of the city.

By piecing together existing 2D images and videos, Scape’s AI-based ‘vision engine’ is able to create 3D models of real-world surroundings at a large scale, which devices can then reference against to pinpoint their location. They recently finished a 3D model of London.

The solution for building a 3D model with a 2D camera was inspired by chickens. You’ve likely noticed that chickens move their head back and forward when they walk, but why is this?

As their field of view in each eye doesn’t overlap, they have no sense of depth perception. By moving their head backwards and forwards, they are able to see objects from slightly different angles, allowing them to map their surroundings out in 3D.

Using the same principle, you can combine 2D images to create a 3D effect.

In order to map out entire cities, Scape applied this technique when analysing and combining the existing images and videos accumulated from sources around the world, allowing them to autonomously assess which images and videos are of the same buildings or features and then using the different angles to measure it in 3D. They explain the process in detail in their ‘Mapping the World’ series which you can read on the Scape.io blog.

Recognising your surroundings

When outdoors in a city, devices are able to interpret their location using a single camera image, which queries against Scape’s maps held in the cloud. Scape’s Visual Positioning System feeds back the exact location and orientation of the device via its API.

The applications for a camera that can recognise it’s surrounding (far more precisely than GPS alone) are extensive. It is essential, for example, that Driverless cars can accurately recognise their surroundings, both for navigation and for safety. Scape’s Visual Positioning System API will allow developers to easily integrate this tech into their software.

Augmenting Reality

Using augmented reality, your phone will also be able to project images onto your surroundings at a much greater scale than previously possible.

For example, if you’re trying to find your way from A to B in the city, a navigation program built with Scape’s SDK would allow you to point your phone camera at your surrounding and be guided by floating arrows. You could check the rating of a restaurant, see which shops are on sale, see buildings that are yet to be constructed, and so much more, all through your phone’s camera.

The big picture

This technology has countless real-world applications for convenience, education, entertainment and advertising. Enabling devices to get a better understanding of their location and their surroundings will be essential with the growth of autonomy, allowing for stronger cooperation between a network of more intelligent and effective autonomous devices.

“Initial use cases in AR span multiple industries; entertainment apps (next-generation Pokemon Go), architecture, utilities, city planning, advertising and navigation. Beyond AR, industrial use cases in Micromobility and Autonomous Vehicles are emerging rapidly. In coming weeks Scape will be publicly releasing its ScapeKit SDK to AR developers in London, with more cities to follow” — Grace Gimson, Chief of Staff.