At the 2017 Mobile World Conference in Barcelona, Spain, the hot topic was surprisingly not the next consumer gadget. Instead, the telecom community was consumed by a connection that did not yet exist – 5G. At the time, it was not clear what 5G would entail, but one year later business cases solidified and the “Global System for Mobile Communications Association predicted 5G connections to reach 1.1 billion by 2025, boosting overall operator revenues to $1.3 trillion”.
The February 2019 Point of Beginning cover story, “Mapping the Future for 5G”, discusses how Land Info Worldwide Mapping “has been providing that essential market intelligence to help support telco companies in their preparation for emerging 5G rollouts”.
According to Land Info president Nick Hubing, “in the telco industry, network coverage is king,” and these companies require accurate and up-to-date mapping for the markets they serve. Land Info has been using a combination of aerial images and LiDAR point cloud data within Trimble eCognition to produce 3D building and tree vectors as well as a 1m resolution land cover classification for metropolitan areas across the U.S.. With this detailed information, telecom companies are able to determine the best position for their 5G sensors, guaranteeing that their signals reach the highest number of users. “5G will require lots of small cells installed on city infrastructure like utility poles, lampposts and buildings to ensure line of sight is maintained,” Hubing says. “The 3D mapping we are providing enables telco companies to readily identify the optimum locations for their infrastructure and accelerate their rollout.”
Chris Lowe, the director of image analysis at Land Info, has been responsible for the development of the 3D city models and brings s great deal of cellular network design experience to the team. When Chris first started planning such networks 20 years ago, the work required a large team to painstakingly digitize old topo maps with a resolution of 30-90 meters. “Today, planning for 5G networks requires the accuracy and detail of 1-meter datasets. In sharp contrast to 20 years ago, however, our OBIA technology allows us to process, classify and map significant volumes of data with a small team” Chris says.
To achieve this accuracy, Chris is using 1m multispectral imagery from the USDA’s National Agriculture Imagery Program (NAIP) as well as DTMs and DSMs derived from LiDAR or SGM (Semi-Global Matching). The data was then fed into eCognition Developer and processed with the rule sets that Chris created to automatically extract the features required for the 5G market.
For this, three rule sets were used. The first focused on tree extraction and was specifically designed to consider tree shadows, incorporating the strong contextual relationship features available with eCognition and unique to OBIA. The second, analysed the building footprints within each AOI – the 5G networks demand a high level of building detail and thus building footprints had to “represent buildings like they are in the real world, with different levels and elevations”. The final rule set refined the combined building and vegetation classifications – Chris designed an algorithm for the software to identify elevated vegetation within a building’s footprint, “erase” the vegetation and then use the nDSM to attribute each building segment with a mean height”.
The final deliverable to the telecom companies were 3D city models complete with vegetation contours, building footprints and the land cover classification.
Land Info has provided these customized 5G maps for more than 30 metropolitan areas in the U.S. to-date and expect demand to grow as the 5G rollout continues in 2019. With the level of automation and standardization that eCognition allows, they are confident that their team can meet the future mapping demands.
“I typically only had a few days to process each AOI, and often, I was processing more than one simultaneously,” Lowe says. “The speed and automation of eCognition allowed me to process 100 square kilometers in about 35 minutes, but with the scalability of the software, I could process 10 times that just as quickly. What’s really impressive is that the datasets come straight out of the software nearly customer ready”.