Google Maps API Maximum Zoom

We have recently been looking at our map created from the Google Maps API’s Maximum Zoom data:
Google Maps API Maximum Zoom – Part 1: Data collection
Google Maps API Maximum Zoom – Part 2: Overview
Google Maps API Maximum Zoom – Part 3: Starting on a more detailed look
Google Maps API Maximum Zoom – Part 4: Continuing the detailed look
Google Maps API Maximum Zoom – Part 5: A detailed look – the Americas

As we have mentioned before, the maximum zoom available in Google Maps varies between 7 and 22. However, if we correct for the magnification at the poles, the minimum becomes 9. Below you can see samples of imagery at some of the zoom levels. We have skipped over a few levels either due to lack of imagery, or because they are too close to see much difference. Note the scale on each image.

We have corrected for the magnification at the poles and then separated out each zoom level into its own overlay, so if say you want to find all Landsat imagery (Zoom level 16) then you can display just that.

To view the overlays in Google Earth, download this KML file

When looking at the highest zoom level our map only picked up very high resolution imagery in Europe. However, this is partly because our survey was not fine enough to pick up small patches of high resolution, such as:

• An image taken by Frank of the X-Prize cup in Las Cruces, New Mexico

• Imagery in Australia captured for Australia Day

• A large number of images in Africa captured as part of National Geographic’s “Africa Megaflyover”

We have recently been looking at our map created from the Google Maps API’s Maximum Zoom data:
Google Maps API Maximum Zoom – Part 1: Data collection
Google Maps API Maximum Zoom – Part 2: Overview
Google Maps API Maximum Zoom – Part 3: Starting on a more detailed look
Google Maps API Maximum Zoom – Part 4: Continuing the detailed look

Today we are continuing with the detailed look, this time at the Americas.

South America

     As we have seen with other areas, South America has two sets of background imagery. Low resolution Landsat imagery is used in the northern half of the continent and the southern countries of Central America. There is also a patch of Landsat imagery in the eastern part of Brazil

     The rest of the continent uses Cnes/Spot Image as the background image.

     High resolution satellite imagery largely matches the population density and is sparsest over the Amazon Rain Forest. We had expected to see a lack of imagery over the Andes due to snow cover, but in fact it seems the Andes have relatively good coverage. The approximate line of the Amazon River can also be seen to have better coverage than surrounding areas.

     Spots of aerial imagery can be seen in a number of Brazilian cities and Cordoba, Argentina. In all the areas we checked, the imagery showed no attribution, suggesting it belongs to Google, probably gathered during its 3D imagery collection. Note that 3D imagery is not included in this data and South America does have quite a lot of 3D imagery.

North America

     Landsat imagery is used as the ‘background’ for northern Canada. The Great Lakes also show up as this resolution, but they have sea floor data (supplied by NOAA), not satellite imagery. Lake of the Woods, a lake in Northern Minnesota, has low resolution imagery attributed to Terrametrics, as is Lake Victoria in Africa.

     Mexico and much of Canada use Cnes/Spot Image as the background imagery. Interestingly it seems to have a northernmost limit of about 60.3°N

     High resolution satellite imagery coverage is poor over Canada, especially towards the North. Mexico has reasonably good satellite imagery coverage.

     Canada and Mexico do have some small patches of aerial imagery. The US is notable for being entirely covered with aerial imagery. Even the lighter shades in the US are low resolution aerial imagery. Although a lot of the imagery was actually gathered by Google, the complete coverage can be attributed to government agencies, which gather aerial imagery and Google has obtained it from them.

     The US has a number of extra dark red patches indicating extra high resolution aerial imagery.

Note that the US, Canada and Mexico all have significant 3D imagery that is not included in this data.

We have recently been looking at our map created from the Google Maps API’s Maximum Zoom data:
Google Maps API Maximum Zoom – Part 1: Data collection
Google Maps API Maximum Zoom – Part 2: Overview
Google Maps API Maximum Zoom – Part 3: Starting on a more detailed look
Today we are continuing with the detailed look at specific regions of the world.

East Asia

     Lake Baikal in Russia, shows up in the data as a yellow streak. It has lake floor data from “Data INTAS Project 99-1669”. As far as we know, this lake and the Great Lakes of North America are the only lakes to have floor data in Google Earth. We had never really noticed this lake before, despite it apparently being the largest freshwater lake (by volume) in the world.

     Landsat imagery can be seen as a ‘background image’ in parts of Tibet, North East China, about half of Mongolia and most of Russia.

     CNES/Spot Image data is used as the ‘background image’ for the remaining parts of East Asia.

     Satellite image coverage is quite good for East Asia and steadily improving. The survey of recent imagery we looked at last month showed significant additions of CNES/Astrum imagery in China and the Koreas.

     Aerial imagery can be found in Taiwan and Japan. As we said in our last post, note that 3D imagery is not shown in this dataset. Japan now has a significant amount of 3D imagery, and in the last few days Hong Kong was added. We are not aware of any aerial imagery in Mainland China or either of the Koreas.

South East Asia

     The tropical forests of the Malay archipelago have very poor satellite imagery coverage. More than half the land is covered by the ‘background’ Landsat imagery.

     The continental countries of South East Asia use CNES/Spot Image as their ‘background imagery’.

     The Malay archipelago has quite a lot of false colour or black and white satellite imagery in ‘historical imagery’, most of which is heavily clouded, suggesting that the reason for the lack of high resolution satellite imagery is due to frequent cloud cover.

     The only aerial imagery in this region seems to be a patch covering Manila in the Philippines.

Australia and New Zealand

     Australia and New Zealand have CNES/Spot Image as their background imagery.

     Australia’s high resolution satellite imagery mostly matches the population density, with rather poor coverage over the interior deserts. New Zealand’s coverage is quite good.

          Both Australia and New Zealand have aerial imagery for major cities and they also have 3D coverage, which is not reflected in this data. The aerial imagery varies slightly in resolution, hence the different shades of red. Our dataset did not pick up the very high resolution patches captured for Australia Day in 2007.

In our next post in this series we will look at the Americas.

We have recently been looking at our map created from the Google Maps API’s Maximum Zoom data:
Google Maps API Maximum Zoom – Part 1: Data collection
Google Maps API Maximum Zoom – Part 2: Overview

Today we are starting to take a more detailed look at specific regions of the world.

Africa

As we have explained in the past, Google Earth has several ‘background imagery’ data sets that it uses when there is no good quality high resolution satellite imagery available. In Africa there are two distinct sets of ‘background imagery’.

     Across the equatorial region of Africa and a curious strip in the east of Libya, the background imagery is low resolution Landsat data, essentially the same as is seen globally when zoomed out. This background imagery shows through most noticeably where there are rain forests. We believe the lack of higher resolution satellite imagery there may be due to the high frequency of cloud cover.

Lake Victoria is an exception and has imagery attributed to TerraMetrics which, according to this page, used to supply the global ‘background image’ for Google Earth.

     For the rest of Africa, the default background imagery is slightly higher resolution satellite imagery from CNES/Spot Image. It is most visible across the Sahara. We believe the absence of higher resolution imagery over the Sahara is a combination of lack of interest due to the low population and difficult photographic conditions. It is notable that much of the high resolution satellite imagery in the region is false colour imagery suggesting that visible light imagery is difficult to capture because of the bright desert sands.

     Most of Africa has reasonably good high resolution satellite imagery coverage supplied by DigitalGlobe and CNES/Astrum. It is fairly easy to visually distinguish between the two in historical imagery, as the CNES/Astrum imagery has a greenish tint to it. Also, the CNES/Astrum imagery tends to be in approximately equal-sided parallelograms, whereas about half the DigitalGlobe imagery is in strips covering one degree of longitude or latitude.

     Only two spots of Aerial imagery were picked up by our map: Bloemfontein and Cape Town, both in South Africa. There is more aerial imagery in South Africa, but our dataset is not high enough resolution to pick it up.

Europe

The obvious band at 60° north is explained in this post.

It is important to note here that the data we are looking at is based on maps created using the Google Maps API. See this page as an example. One significant difference between Google Maps created this way is that they do not show 3D imagery. Europe has a significant amount of 3D imagery, so what we see in the Google Maps API generated map may be quite different from what can be seen on the standard Google Maps website or in Google Earth.

     As with Africa, Europe has two sets of ‘background imagery’. Low resolution Landsat data can be seen in Ireland, Austria and the northern parts of Norway, Sweden and Finland.

     The second slightly higher resolution set of ‘background imagery’ comes from CNES/Spot Image.

                 Most of Western Europe is a patchwork of aerial imagery of various resolutions from a wide variety of suppliers, including a significant amount collected by Google itself. There is a distinct transition to satellite imagery over Eastern Europe, more obvious in ‘historical imagery’ than in the ‘maximum zoom’ data. Poland and Romania do have some aerial imagery and the country of Montenegro stands out as having its own patch of aerial imagery almost exactly matching its borders. Northern Scandinavia and Ireland are somewhat lacking in aerial imagery.

The Middle East and Central Asia

     Landsat imagery is used as a ‘background image’ for Saudi Arabia, most of Russia, and parts of Kyrgyzstan

     The rest of this region uses Cnes/Spot Image imagery as the ‘background image’. Iraq and Afghanistan really stand out because they have not had any imagery updates for several years. We believe this is deliberate censorship due to the wars in those countries. The sparsity of high resolution satellite imagery in Saudi Arabia and northern Asia is probably, as with the Sahara, a combination of lack of interest due to the low population and difficult photographic conditions. Saudi Arabia has bright sand, and northern Asia has snow cover and poor light for much of the year.

     Pakistan and India have remarkably good coverage with high resolution satellite imagery. Nepal, Bhutan and Bangladesh are also pretty well covered.

There appears to be no aerial imagery in this region. We believe some countries have not allowed Google to gather aerial imagery and for others there may simply be a lack of local suppliers.

We will continue with further regions of the globe in a later post.

Yesterday we showed you a map of the maximum available zoom level found in Google Maps. If you don’t already have it, download this KML file to view it in Google Earth.

Today we are looking at general features of the data.

Key
First of all, we need to know what the colours mean. The colours represent the various available zoom levels in Google Maps, starting with low resolution at 7 to high resolution at 22. This can be roughly interpreted as follows:

	Low resolution ocean floor
	High resolution ocean floor
	Coastal areas
	Very low resolution satellite imagery (Landsat background imagery)
	Low resolution satellite imagery (CNES/Spot Image background imagery)
	High resolution satellite imagery (DigitalGlobe or CNES Astrium)
	Aerial Imagery (higher resolution than satellite imagery)
	Exceptionally high resolution Aerial Imagery

Google Maps via API
To better understand Google Maps Zoom levels or to explore the data for a given location, you can use this page, which shows a full screen Google Map using the Google Maps API. It behaves a little differently from the standard Google Maps website. The standard Google Maps website restricts how far you can zoom in based on the imagery available, however, it always shows imagery. The above map, obtained via the API, lets you zoom in beyond the maximum prefered amount, and when you do it displays map tiles labelled “Sorry, we have no imagery here”.

Poles
You will notice in our map that there are noticeable bands in the data towards the poles at 60° and 75° latitude, both north and south. We believe these are an artifact of the Google Maps projection, which spreads out the poles, thus magnifying any imagery towards the poles and in consequence requiring less zoom in order to see a given resolution of imagery.

In our data collection we only went 80° north and south. Google Maps itself only goes to 85° north and south as a consequence of its map projection.

Google Earth has noticeable bands in the actual imagery at 80° north across Greenland and about 82.6° south in Antarctica. These are actual changes in the imagery datasets. The absence of high resolution near the poles may be due to the orbits of the imaging satellites or ‘sun angle constraints’.

Oceans
The spider web of tracks across the oceans noticeable in our data reflect the paths of ships equipped with sonar for mapping the ocean floor. The tracks of higher resolution imagery are clearly visible in Google Earth and have in the past been mistaken for Atlantis or an alien base.

In our next post in this series we will be looking in a bit more detail at the various types of imagery found over land.