August 16th, 2016, Published in Articles: EE Publishers, Articles: PositionIT
by Emmanuel Stefanakis, University of New Brunswick, Canada
Geocoding is the process of converting a street address into a physical location that can be described with a pair of geographic coordinates. It is estimated that over 40% of the world population is physically disconnected due to the lack of a street address.
Even when street addresses are available, they are very often unable to describe the location. For example, locations inside parks or large facilities (e.g. stadiums or hospitals with multiple entrances) may be hundreds of metres away from the nearest address. The use of directions (such as “behind the main building find a storehouse; deliver the package at the right door facing the park”) instead of an address has become common practice. However, this is usually ambiguous as it relies on local knowledge, and it cannot be interpreted automatically.
Fig. 1: UNB Campus. Three locations in the Faculty of Engineering building. The street address to all of them is that of the Main Entrance: 15 Dineen Drive, Fredericton, NB.
The wide spread of smartphones and mobile devices on one hand, and extended internet accessibility on the other, have brought back the problem of geocoding in geomatics research and development. A geographic point location described by numbers (coordinates) – the geographic latitude and longitude as well as the height or depth – can be found more easily than ever by most of the users (as the smartphone becomes default equipment).
ISO 6709:2008 is the geocoding system that offers a standardised representation of geographic point location by coordinates [1]. This system may describe any location on the earth’s surface and the surrounding area at any resolution or zoom level.
However, it is argued that geographic coordinates are difficult to remember. In addition, a pair of latitude, longitude coordinates refers to a specific point in space regardless of the number of decimal digits used to describe them. By truncating the decimal digits of geographic coordinates, a new location is defined. This way, geographic coordinates fail to explicitly define an area in space with a size that varies depending on the resolution.
Over the last couple of years there has been a significant interest in the development of alternative location encoding systems to better support specific geomatics applications, including navigation in space (directions). Two popular systems, introduced recently, are: Google’s Open Location Code [2] and What3words [3]. Those systems are added to an already long list of other geocoding systems [4], such as Geohash, Geotude, C-Squares, MapCode, Open Post Code, WMO squares, and UTM grid, to name a few.
A location encoding system (also known as a geocoding system) is a scheme that assigns systematic alphanumeric labels to geographic locations or entities. Location encoding systems can be grouped into two categories:
Fig. 2: Open Location Code encoding. The three locations in the Faculty of Engineering building (Fig. 1) are contained in areas assigned unique ten-character codes. Each area corresponds to approximately 14 m x 14 m.
In the latter category, the tiles are encoded using an algorithm that calculates the alphanumeric strings and avoids recognisable words (e.g. Open Location Code), or the tiles are assigned one or more recognisable words from the English or other dictionary (e.g. What3words).
In addition, some geocoding systems support a multi-resolution scheme and encode smaller or larger tiles (areas) with more or less digits in a code (e.g., Geohash, Open Location Code), while others follow a constant tile size and equal-sized codes (e.g. What3words, WMO squares, UTM grid).
Existing location encoding systems have been designed to support specific needs, hence their attributes vary accordingly. Location encoding systems designed to support navigation (directions), such as Open Location Code and What3words, are able to describe any location on earth with a short code. Users can enter this code in their phone, tablet, or laptop, and get the exact location. They do not even need to be online. An application running on their devices can support navigation to the destination.
The desired attributes for a location encoding system to support navigation are [5]:
Fig. 3: What3words encoding. Grid cells are 3 x 3 m large. The three locations in the Faculty of Engineering building (Fig. 1) are contained in cells assigned unique three-word codes.
Each location encoding system meets these attributes in higher or lower degree, and no system can outperform the others. Depending on the application needs, a particular system is preferable over another. For example, Open Location Code supports multi-resolution representation and the extraction of spatial relations between two codes, while What3words does not (all tiles are equal-sized: 3 m x 3 m and two codes cannot be compared to extract spatial relations). On the other hand, What3words provides codes that are easy to memorise; it can also support multiple languages. Open Location Code uses Latin characters and as a result it fails to be language or culturally independent as claimed [6]. Neither system considers the third dimension; this is certainly an extension to expect in the future. Overall, geographic coordinates are too hard to remember, but replacing them has implications.
Acknowledgement
This article was previously published in GoGeomatics Canada on 11 March 2016 and has been republished here with the author’s permission.
References
[1] ISO 6709:2008. www.iso.org/iso/catalogue_detail.htm?csnumber=39242 [Visited on Feb. 19, 2016]
[2] Open Location Code. http://openlocationcode.com/ [Visited on Feb. 19, 2016]
[3] What3words. https://what3words.com/ [Visited on Feb. 19, 2016]
[4] List of Geocoding Systems, Wikipedia. https://en.wikipedia.org/wiki/List_of_geocoding_systems [Visited on Feb. 19, 2016]
[5] D Rinckes: An Evaluation of Location Encoding Systems. 2015. https://github.com/google/open-location-code/blob/master/docs/comparison.adoc [Visited on Feb. 19, 2016]
[6] G Rhind. How Google could improve Open Location Codes. 2015. http://resources.pcapredict.com/index.php/google-open-location-codes/ [Visited on Feb. 19, 2016]
Contact Emmanuel Stefanakis, University of New Brunswick, emmanuel.stefanakis@gmail.com