Objectives of the service
Image credit: Proekspert AS
Growing global population has increased the demand for open-air sports and leisure facilities like football- and rugby fields, and public playgrounds. However, the costs to build and maintain these facilities are also increasing. As a cost-saving measure, more facilities with artificial turf are being built. To plan better investments and optimise maintenance costs, up-to-date information is needed about football fields and their condition. Currently, such information is not easily available and not obtainable at a reasonable cost.
The objective of this feasibility study is to investigate whether it is technically and commercially feasible to obtain this information by using satellite image data and analyzing it with the help of machine learning tools and techniques. Both components, technical- and commercial feasibility, will be validated in the feasibility study.
We build an application that enables basic reports, where information on artificial turf fields may be found with a few clicks at a reasonable cost.
Users and their needs
Manufacturing, construction, and maintenance companies, as well as local governments, have a need for a quick and efficient overview of artificial turf fields in their area, their number and condition.
Good reasons to possess this information:
- To learn about market opportunities;
- To make medium- to long-term investment plans for new fields or field reconstruction;
- To validate and analyse whether demographic trends in an area support the construction of more sports and leisure facilities.
- To determine if environmental regulations allow the construction of planned facilities;
- To monitor field maintenance.
To prepare for the project, we interviewed several organisations involved in sales and reconstruction of artificial turf. Those interviewed said obtaining a report about artificial turf fields in a geographic area by clicking on a spot on a map would be highly beneficial. With this simple tool, users could answer questions such as “What are the fields we were not aware of?” and “Where are fields in poor condition which need renewal or replacement?”
The result would be business intelligence at a qualitatively new level, enabling the fine tuning of medium- and long-term R&D, marketing, and financial strategies.
The application may be used globally.
Service/ system concept
The service uses satellite images which are analysed with the help machine learning algorithms to obtain a list of identified fields covered with artificial turf. A user interface enables map access when the user selects a geographical spot. By clicking on a map location, the customer activates the application to obtain images from that area and apply trained machine learning algorithms to find relevant fields. The application is currently designed to locate football fields, but in the latter stages of product development, fields with other uses will be added to the menu.
As a result of the analysis, the user will receive a list of links to football field locations on the map. The customer may click on the link and continue their investigation into that specific field.
Space Added Value
To identify football fields (Level 1) we will start with Sentinel-2 high-resolution multi-spectral imaging data with spectral bands at 10 m spatial resolution. If the detection and identification accuracy is lower than expected, we will proceed using panchromatic images from Spot-6 satellites with 1.5 m spatial resolution.
For the purposes of the current technical feasibility study, we will use archival images to reduce the cost of the service and build an understanding of the solution’s commercial viability.
A proof-of-concept is under development. As of today we are able to identify semi-automaticaly and with high probability football fields from the satellite images as well as identify which one is covered with natural grass and which one with the artificiall turf.
Image credit : Proekspert