Reason
It rains more irregularly and harder than before. This makes it important that we use rainwater wisely. Surface paving plays an important role in preventing flooding and drought. However, paving and other surfacing are constantly changing.
At NEO, we identified the need to better anticipate flooding and drought with better information. In a nutshell, this involves information on the surface hardening of the Netherlands that is uniform nationwide, but also detailed and up-to-date. To have a correct, complete and above all up-to-date data supply uniform, country-wide and up-to-date. The need and number of questions on this subject prompted us to invest in a service that uses geo-information to provide insight into surface paving and makes it easy and fast to consult this information. The result is a digital service in which public space and private properties are divided into a number of surface paving types. We call the service the Pavement Monitor.
The company
NEO is a geo-ICT company with over 30 employees that provides services based on Earth observation data and other big data. The company will be in its 25th year in 2021. We focus on monitoring changes in our living environment and in this we are unique in the world.
By 2021, we will not only monitor paving, but also do so for all buildings, roads, water courses, trees, fields, nature areas, solar panels, etc. throughout the Netherlands. We use satellite images, aerial photos and lidar data as well as data from the 'crowd' and other Big Open Data. Our patented technology provides automated change signals with unique high completeness, correctness and timeliness. With the observed changes, we build information services that we unlock via our API (Application Programming Interface = our digital waiter for data). Examples of these new services are in there in crop management, map registrations such as those in BAG, BGT and WOZ, in tree maintenance, pipeline management, building inspections and control, invasive exotics, etc.
The pavement monitor
In the Pavement Monitor, we have divided every piece of surface in the Netherlands into one of the following classes:
Closed pavement
Open pavement
Unpaved
Water
Building
The classes are assigned to the objects at ground level in the most up-to-date version of the Basisregistratie Grootschalige Topografie (BGT). For surfaces for which the BGT does not provide pavement information such as business premises, gardens and yards, among others, the pavement class is assigned per area of 50*50 cm from analysis of aerial photographs and satellite images.
The data has a level of detail at least equivalent to that of the BGT, but also includes the information for farmyards and business parks, i.e. those parts of the map where the BGT does not contain information. Moreover, we permanently process all changes in pavement, etc. using Open Data and images.
The Pavement Monitor is offered as a digital service to which you can subscribe and which allows you to provide services to third parties, or whose data you can make available to third parties working with the data on your behalf. The data are offered via an API.
Data quality
The Pavement Monitor covers the entire land surface of the Netherlands. In 2021, the Pavement Monitor will be permanently updated with new data. In principle, the BGT lags behind reality by a maximum of 18 months. Changes processed in the BGT are adopted in the Pavement Monitor immediately after publication. Changes in the surfacing of business premises, yards, gardens and building sites are always processed in the Pavement Monitor within a few months of the change occurring and usually within a few months.
The Pavement Monitor follows the geometry of the BGT and is just as accurate.
Technology and collection
The data are updated with aerial photos and satellite images and the BGT. The workflow is designed so that the entire updating process is automated. AI (deep learning algorithms) plays an important role in this process. Change signals from images and BGT lead to adjustments in the Pavement Monitor. In addition, the algorithms are naturally checked and better trained, especially for strange situations. The automated solution for a strange change is that it is parked until clarity is obtained with the next image (think of construction sites).
Other data sources can also help improve classification, especially in exceptional situations. We carry out a first step in this with data from the Tree Register. Tree crowns act as 'interferers' for surface visibility on aerial or satellite images when they are in leaf. Using image information around tree crowns only from leafless seasons prevents misclassification.
Use
The pavement monitor is a digital map with up-to-date information and those interested in it can access the map free of charge at the web address The pavement monitor (neo.co.uk This will allow users to gain their own insights into specific locations, etc. More importantly, however, specialists can use the data on the API as input for calculations on flooding and drought, water inflow for sewers, water availability for urban trees, dynamics of urban groundwater, etc. A licence fee is charged for using the API. The licence allows easy transmission of information.
A number of uses in which the data is of value:
Monitoring urban greening, e.g. as part of Operation Steenbreek or research/policies on climate adaptation or heat stress;
Run-off modelling in the context of sewer loading, identifying bottlenecks during extreme rainfall events and determining solutions such as strategic placement of green strips or permeable pavements, among others;
Excavation/ subsurface management, e.g. when laying fibre optics or maintaining pipes;
Subsidising private green space or its control.
Future
NEO is working on the further development of the Pavement Monitor. Additional information will soon be added regarding slope of map objects, detailing of pavement type, etc. We will soon publish a viewer on our website www.neo.nl in which the data can be viewed country-wide. We are always open to good suggestions from users or interested parties, so if you have any thoughts on this or are curious about the data, please let us know.