Objectives of the service
The proposed Building Stock Monitor Service is a service to aid municipalities with monitoring the entire building stock within their geographical borders.
Presently, municipal map information and accompanying building polygon data are updated with a low frequency, depending on the budgets available. Since performing fly-bys is expensive, the update rate can be up to several years.
By using satellite data, the proposed service will be able to give the municipalities weekly updated data for their community. The updated information makes it possible for the municipalities to work more efficiently.
For instance, the service will enable municipalities to closely monitor authorized building projects and detect whether they comply with the given permit or not or detect illicit building activity immediately and prevent the activity from developing further.
Furthermore, the municipalities can address the usual concerns of basic infrastructure, security, environment and social and economic issues with accurate data that is key to all modern urban and rural planning.
Users and their needs
The service’s target users are case officers, planners and operative staff of a municipality. Typically, their daily operations consist of evaluating and following up on building permit applications (case officers), monitoring illicit building construction (operative staff) or perform urban or rural planning (planners).
Case officers need to know whether the construction is progressing as planned and does not violate what was granted as per the application, and also need to approve the final building for tenancy. By monitoring this via remote sensing, the need for field inspection is minimized, and the work process becomes more effective.
The operative staff needs to know whether citizens raise buildings illicitly. Such activity might be a risk to safety and security, violate social and environmental regulations, and can cause unnecessary disputes that might have costly consequences.
Planners might want to consult historical activity in areas when regulating different areas for different purposes.
The service will first be launched in the Norwegian market
Service/ system concept
This project has investigated the feasibility of a monitoring service for delivering change detection on buildings. More specifically, whenever there is a change in building stock, e.g. a new building, a changed building (usually augmented) or demolished building, the users receive alerts on where the change occurred, the size of the change and the time it happened.
The various user groups of the municipalities then use the information according to their field of operation: case officers consult and compare their list of assigned permit applications and check whether construction is going as planned, operative staff checks if the buildings are illicit or not and perform field inspections if necessary.
The overall concept of the service is depicted below.
(The image above is an example.)
Space Added Value
This project has investigated whether the proposed service could utilize SAR data from Sentinel-1 and optical data from Sentinel 2 to perform change detection using machine learning and deep neural networks. Change detection is performed in both domains.
Current methodology for updating building polygons and official cadastre information typically involves performing fly-bys at intervals spanning from 1.5 to 5 years, depending on individual municipalities’ budgets. The data-quality gradually degrades and the back-log becomes increasingly difficult to manage, the longer the update-interval.
The revisit time of about 6-8 days for the Sentinels over Norway yields an unprecedented update frequency and will help managing backlog and increase data-quality.
The image below shows before and after imagery from Sentinel-1 and the change map produced by the machine learning algorithms investigated in the study.
(The image above is an example.)
The feasibility study’s first and foremost objective was to investigate whether change detection was possible using Sentinel data and machine learning.
By the completion of the study we have shown that this is indeed feasible, and that municipalities that have participated in the study will indeed have use of it.
It is the intention to develop the proposed service and bring it to the market as fully operational service via a Demonstration project.