A Seminar Series in GIS, Remote Sensing, and GeoInformatics organised by the Edinburgh Earth Observatory in conjunction with the AGI-Scotland
Friday 16 October 2020
Friday 16th October saw the second installment in this autumn’s EEO-AGI(S) seminar series, when we welcomed David Cameron-Mackintosh to talk about “Modern Surveying: Tech and Technique”. David is a Sales Director at Survey Solutions Scotland, one of the largest survey equipment suppliers in the land, and his whistle-stop tour covered developments in the areas of both discrete and mass data collection technology.
Discrete data collection refers to techniques such as total stations, which are becoming increasingly automated and remotely-operated, as well as GNSS, known to most outside the profession as GPS devices. While the cm-accuracy that is possible with GNSS instruments has not changed significantly in recent years, the real improvements have come in the areas of portability and signal availability. Permanent base station networks, smaller devices, and integration of inertial measurement units (IMUs) have all contributed to these recent advances.
Significant gains have been made also in mass data collection techniques, such as laser/LiDAR and photogrammetry. Laser systems can now be mounted on everything from trains to cars to backpacks, with acquisition rates reaching 2 million points per second in some cases – talk about big data! Other developments have seen range increases and automated registration, allowing near-real time review of data in the field. LiDAR (used here in the context of aircraft-hosted laser systems) has experienced similar technological advancements, such as integration of GNSS and IMU technology.
David ended his talk with a demonstration of the cutting-edge software that is available from companies such as Trimble, which allows integration of all of these technologies and datasets.
Euan Mitchell, MSc Geographical Information Science
Friday 2 October 2020
The EEO–AGI(S) seminar series returned on Friday, 2nd October at a new time of 5 pm, with a talk from Prof. Michael Peterson, Professor at the University of Nebraska at Omaha and author of several books, including Mapping in the Cloud. As with seemingly everything these days in our new post-Covid world, Prof Peterson’s engaging presentation on “Evaluating Feature Density on Large Scale Online Maps” was delivered on Zoom.
Multi-scale, pannable maps are ubiquitous in modern society, thanks to the proliferation of smartphones and applications such as Google Maps, Bing Maps, and several others. All online map providers use the same system of 256 x 256 pixel tiles, delivered seamlessly at ever increasing zoom (not Zoom!) levels, even down to use of the same individual tile boundaries. The only difference, therefore, in their end products is the underlying database of vector data and the rendering of those data in map form. The purpose of Michael’s talk was to evaluate three major mapping services by comparing feature and label density. The mapping services evaluated were: i) Google Maps – the most popular such service; ii) Microsoft’s Bing Maps – who rely on data from the Dutch companies Here and Tom Tom; and iii) MapBox – a California-based map API provider that relies exclusively on volunteered OpenStreetMap (OSM) data for their products.
The comparison process entails a side-by-side rendering of maps from different providers, at an extremely large scale, centered on a randomly chosen location in North America, Europe, or Sub-Saharan Africa. For each map pair that returns features (which is the exception rather than the rule at this zoom level, where each map represents an area of approximately 150 x 100 m) the ‘winner’ is the map that displays the greatest number of features, unless a ‘tie’ is declared. The process was repeated until 100 maps had been evaluated for each pair of providers in each region. The code that generates the maps can be found on the website accompanying the Mapping in the Cloud book.
In North America and Europe, Google Maps was the clear winner over Bing Maps. In Africa, however, Bing significantly outperformed Google, which is the default on the Android devices that enjoy an 87% market share in the African cell phone market. The feature density of MapBox maps, using OSM data, was the lowest in all three areas. This highlights that although OSM data compare favourably with national agency datasets, such as those of Ordnance Survey, in terms of spatial accuracy, data coverage is much more heterogeneous, being more complete in urban and affluent areas.
Prof. Peterson wrapped up his presentation with some areas for future exploration, such as expanding the comparison to other regions of the world, and evaluation of attributes such as label clarity or rendering quality, before posing the question ‘How do we get to an accurate map of the world?’.