Aerial imagery contains valuable information
Aerial imagery is well suited for surveying a building or an area quickly. But, what do you do if you want to inspect thousands of buildings or a whole country? Doing it manually isn't an option!
Readaar mines data from aerial photographs, satellite images and LiDAR data. We use object recognition, machine learning, change detection and dense image matching to translate imagery or point clouds into actionable knowledge and CAD-files.
A house says a lot about its owners. For example, the owners of solar panels constitute an interesting target group for a broad range of products. We detect and map them on an address level. This is interesting if your product is related to solar panels, but also if you’re selling solar panels because people buy them only once. Besides solar panels, we map a whole range of other housing characteristics.
Grid operators and energy suppliers
Solar energy use is growing rapidly .This is good news, but there are challenges as well. One significant challenge, for example, is in rural areas where there is a risk of grid overload (small grid capacity vs big installations). Energy suppliers have to balance demand and production and therefore need to know the total amount of solar energy produced. Readaar is able to offer insights about photovoltaic installations at the address level: installed capacity, orientation and inclination.
Home insurance companies
Dormers, extensions, solar panels, etc. raise the construction costs of a building. Most of the time, it is mandatory to report these kinds of changes to the insurance company. However not everybody does this. Readaar helps keep your portfolio up to date, by providing yearly updates of all building changes.
To detect and map solar panels, we use a combination of different techniques such as object recognition and change detection. With these techniques we derive countrywide information such as year of construction, surface area, installed capacity, orientation and inclination.
We map asbestos roofs by determining 100 different characteristics from different sources, we use these to calculate the chance a roof contains asbestos. We are currently executing different pilots with different municipalities. The data will be available early 2017.
Tuning to your particular problem, we develop algorithms that can translate 3D-data into: heights, surfaces and volumes of buildings and area (DSM/DTM). Depending on the requirements, we use LiDAR-data or dense image matching on stereo imagery.
Modifications are typically either replacement, extension or removal. All these can be identified and mapped using change detection. For replacement, we compare imagery of different years; for volume alteration, we use LiDAR-data or dense image matching based on stereo imagery.
Sven can do anything with aerial photographs, satellite images and LiDAR data. The bigger the challenge, the better!
Jean Michel Renders
Jean Michel can do anything with data, he combines a Phd in machine learning with corporate experience in data mining and artificial intelligence
Matthijs van Til
Matthijs is an entrepreneur who loves technology, innovation and sustainability.
Erik ensures that our organization is running smoothly so that we deliver on time and according to specifications.
Jasmijn is responsible for quality control, she makes sure everything is checked before we sent it to customers.
Martijn is working on his final thesis about the improved detection of roof segments.
Carlos Camps Pons
Carlos studies aerospace engineering and works on the detection of roofmateriaals.
Jeroen labels asbestos roofs, he recognizes them from a distance.
Michael has a sharp eye for details and uses that to label asbestos roofs.