Tracking the progress of housing projects through satellite imagery

Scottish space data specialist, Astrosat, combines innovation in the space sector with information available on earth to create practical applications with real-world impact. The firm uses information and images taken from satellites to add value to the existing datasets gained here on Earth. The result is comprehensive, digestible material presented in a visual form to help make big changes.

To support the company’s development of a new software tool that could automate the analysis of housing construction progress using remote sensing images, Astrosat signed up to the SPRINT programme to access expertise from the University of Leicester. This collaboration was aimed at developing a machine learning algorithm that would analyse Earth observation and other geospatial data. This would then enable local authorities to support their housing delivery monitoring needs and more effectively forward plan public services.

Using data from space to monitor house building

The SPRINT project supported Astrosat’s collaboration with the London Borough of Waltham Forest.

The firm was one of five successful suppliers chosen by Waltham Forest Council for the ‘GovTech Catalyst Challenge: Accelerating the building of housing using geospatial intelligence’ tender. The Council was investigating how technology can be used to monitor developments of all sizes and types, to build an accurate picture of development status from permission to completion and occupation. The successful project would develop a tool to help planners make more informed decisions and improve monitoring processes.

The overall programme of the GovTech challenge was delivered in two phases. The first phase involved research and development contracts being awarded to demonstrate technical feasibility of the proposed solution.

The second phase encompassed the awarding of two R&D contracts to businesses chosen from the successful Phase 1 applicants – Astrosat is coming to the close of Phase 2 now. Here, Astrosat turned to the use of satellites to detect stages of construction, from commencement to completion.

In tracking the progress of housing projects around the borough, Waltham Forest Council wanted to determine if construction had taken place and identify the gap between planning permission and when construction started.

To achieve this, Astrosat leveraged the University of Leicester’s work on image processing, change detection and machine learning methods to process high resolution Synthetic Aperture Radar (SAR) imagery, replicated it and improved its accuracy.

Waltham Forest Council spends an estimated £750,000 a year tracking the progress of housing projects around the borough and hoped to find a cheaper, more streamlined way to do so. It cost the council between £700 to £800 per month just for satellite data which did not deliver sufficient benefits for a public authority body to merit its use. As a result, the council looked at SAR data, which pivoted from the University of Leicester, as a more cost-effective option.

Benefits achieved through SPRINT’S support include:

• Won phase 2 of GovTech project on the back of SPRINT, worth £0.5 million

• Reduced product development time by at least six months, thus shortening the time it takes to generate a market ready product

• Two more companies have contacted the principal investigator since the completion of the project

There’s a ‘space solution for every earth-bound problem’

Mike Payne, Head of Product Management at Astrosat said: “Local authorities need more visibility of construction projects and this project with Waltham Forest has challenged our vision that there is a ‘space solution for every earth-bound problem’. We could have tried to develop the complete solution in-house but would have needed a significant R&D budget. The SPRINT funding linked us to the University of Leicester to speed up product development and enabled us to commercialise the output of that relationship.

“The SPRINT project has offered us the flexibility to commercialise the solution with other councils. Furthermore, our team of in-house data scientists have been able to apply technology developed in an academic context to real-life applications.”

A proven tool for government agencies

Zhou (Joe) Huiyu, Professor of Machine Learning at the University of Leicester added: “While this project with Astrosat may seem small, it has had a huge impact. It has enabled us to work closely with Astrosat to develop a software tool that combines AI technology to monitor land use and that which has been surveyed by the city council.

“The project has developed two major tools, one which uses software to detect buildings, and another detects whether a new house has been set up from 0 months to three and six months.

“Astrosat applied the technology for the different applications within the local authority framework projects. The proven capability of using remote sensing images to detect or track the development of housing projects will be useful to many government agencies.”