Apple Maps , for example, showed the Australian city of Mildura as if it were inside a national park, where temperatures can reach 46 degrees Celsius, when it was actually 70km at northeast.There were also cases of islands that appeared twice, hospitals and supermarkets that mixed and other errors like that.
Cupertino had decided to separate from Google and create his own maps .Since then six years have passed and Apple has hired engineers and used multiple tools to improve its system mapping, even with the help of special vehicles.The effort has not been in vain and today the maps of the Apple have reached a good level.In addition, all this work will probably be rewarded with the parallel development of its autonomous driving system However, Google Maps is still far away.
Apple vs.Google
In a very detailed article called ' Google Maps's Moat' , the well-known American cartographer Justin O'Beirne illustrates the evolution of Apple and Google’s cartographic data over time.In this comparison, the expert explains that the competitive advantage of Google lies in the time available to improve your system .«Big G» has collected so much data, in so many areas, that Apple can't compete right now.
As a reference, Google Maps was launched in 2005 , and Street View was added two years later. Apple , on the other hand, launched its application on 2012 .In this industry, seven years of difference is an abyss, since they have allowed Google to have more time to capture detailed data from the various areas of the planet.
The analysis shows that Google has been constantly adding details to its maps , with a level of precision that extends to tool stores, park shelters and garages.Apple Maps, on the other hand, It offers much less detail in most of the areas.
O'Beirne explains that Google creates this data using its own satellite images, while Apple probably takes its data from third-party providers such as TomTom.By being able to internally create its own library of information and mapping data, Google has been able to generate your own metadata, which is used to extrapolate even more information.It is a competitive advantage that continues to grow driven by itself.
Greater accuracy in Google maps
According to O'Beirne, Apple has lost some competitive advantages , giving the competitor the opportunity to have a much more detailed mapping that, in the face of the user, translates into more instructions precise and immediate .The qualitative difference is found not so much in big cities, but rather in extra-urban areas.
One of the deficiencies detected during the analysis is the treatment of the footprint of the buildings (including houses, commercial warehouses and others), which seems particularly poor, if not completely absent, in Apple maps.The cartographer has noticed that Google can propose the silhouettes of buildings including rural areas that are located in that 1% of U.S.territory not covered by Google Street View (or where the image capture service has been more than 9 years).
The degree of detail is acceptable even in the presence of small housing complexes and also the green areas seem to receive a very approximate treatment in the case of Apple Maps, unlike Google that represents them correctly.
By zooming from above in satellite mode , Google Maps allows you to see the roofs of buildings with a resolution and details of good quality, especially when compared to the images of 10 years ago.Developing the 3D of the buildings, Google has invested concrete efforts and technology since 2012, when the rendering of a thousand buildings passed in a few months to 25 million.
But technological progress does not stop there.Google Maps has also been able to partially resolve the offer of a series of fundamental information for the user or, such as the identification of so-called commercial corridors , those routes of particular interest for free time and shopping.
Google retrieves information through an algorithm capable of intercepting restaurants, bars and stores .I started doing it in July 2016 with the introduction of a strategy to show the AoI ( area of interest ).To correctly assign this information, there would be 80,000 million processed data, to which is added a technique of computer vision capable of assigning the toponymy of automatic way.
To create good maps , excellent drawing skills are not enough, more and more data is needed to process.And Google has no rivals in this. machine learning It is essential both for the management of names and for the recovery of 3D models from images.This is what Google calls the Ground Truth project, started in 2006.For Apple all this even is far away.
Images | iStock-Justinobeirne.com
Comments
Post a Comment