Notice that the filtering and compositing are done in Google Earth Engine servers and it is, therefore, a very quick process. The download_data function in line 23 will search all the images available in the interval and compute the median of the three images with the lowest cloud cover fraction, as defined by the parameter use_least_cloudy that is set to the value of 3. The time interval selected starts on 1 of March and ends on 25 of April so that the composite will give a recent overview of the construction progress as of the date this story is being written. The region selected in this example corresponds to the location of one of Tesla’s Gigafactory that is in construction near Berlin, Germany. The code above shows, as an example, how you can download Sentinel-2 data (line 11) bands Red, Green and Blue (line 12) for a region delimited by the bounding box coordinates defined in lines 5–8. First, there is the web editor where you can create your scripts that run in google servers and send back the result to your browser where it is displayed in a dynamic map (see image below as an example).Įxample of code to download Sentinel-2 median composites of least cloudy images for a selected region and time period. There are several ways to use Google Earth Engine.
Google Earth Engine provides this data and quick processing capabilities at a global scale. And remember this is just a 100 x 100 km region. Now imagine you want to download one such image every 5 days (to get a time series of the land change) for a period of 1 year and you get nearly 45 GB of data.
If three channels are needed it’s already 600 MB of data. If the image is stored with 16-bits precision without compression that’s 1,600,000,000 bits or 200 MB for each channel. Sentinel-2 data is a good example as it consists of multi-spectral satellite images with up to 10-meter spatial resolution and an orbit revisit period of 5 days considering both Sentinel-2A and Sentinel-2B satellites.Īs an example, if you are studying a region of 100 x 100 km and your data spatial resolution is 10 meters, your image size will be 10,000 x 10,000 pixels. In other words, it is a service freely available that combines a large set of global datasets that often would be impossible to handle without a super-computer.
“Google Earth Engine combines a multi-petabyte catalog of satellite imagery and geospatial datasets with planetary-scale analysis capabilities and makes it available for scientists, researchers, and developers to detect changes, map trends, and quantify differences on the Earth’s surface.” According to the main webpage of Google Earth Engine:
0 kommentar(er)
