Solar Monkey will soon enter the Spanish market. In the Netherlands, many installers already use the software of market leader Solar Monkey to design solar panel installations. To make detailed yield calculations
, Solar Monkey uses height maps. With these maps, it can be calculated how much sunlight will reach a solar panel and a highly accurate estimate of the yield of every solar panel can be made. In the Netherlands, the ‘Waterschappen’ provide these elevation maps (AHN
) as open data. This data is collected from aircraft using laser systems (LiDAR
To be able to offer a similar product in Spain, Solar Monkey also needs a height map there. Unfortunately, LiDAR data is not widely available in Spain. Solar Monkey, therefore, approached Readar
for alternatives. Readar has developed a product to generate elevation maps using aerial images. It uses stereo reconstruction: because every point on the ground is visible in at least two photos, the height can be determined. The process of reconstructing the height of every pixel in the photos is called “dense matching”.
Readar uses a Convolutional Neural Network (CNN; Deep Learning) for dense matching. The big advantage over traditional Dense Matching techniques is that less aerial photos are needed to achieve a good height map. Traditional techniques are preferably applied to photos where each point is visible in 4 to 5 photos (80% forward overlap). These aerial images must be collected especially for this purpose. When matching through a CNN, we only need to see each point on the ground in just two photos (60% forward overlap). This allows the technique to be applied to already available aerial photos.
Readar can generate current elevation maps throughout Europe. Interested? Please contact Matthijs van Til, 0 6 54 914 858 or email@example.com