Google Earth Engine- An Introduction
A prologue to motivate the exploration of Google Earth Engine
1. Reminiscences of the past
Back in the school days around two decades ago, when I heard about Google Earth, I thought of it as an object made for fun. My friends and I used it to zoom in the places of our interest such as our homes, schools,
play fields etc. Never did it occur to us that it is so much more than just plain aerial photographs of the earth when looked from the outer space. Well… the innocent young minds cannot be blamed as documentation and community discussions were scant in those days and so was access to internet. Internet cafes were the only place where you could access internet.1MBPS was the gold standard in those times. Those were the days of hotmail and yahoomail! There was no YouTube. 40 GB hard disks were considered to be sufficient for personal needs.
Fast forward 20 years later, we now have state of the art computation technologies. We now talk about Big Data, Artificial Intelligence, Machine Learning, Internet of Things etc. which were incubating in the realms of science fiction. With the advancements in data storage technology and reduction in costs, it is now possible to store massive amounts of data. To put things into perspective, 500 hours of videos are being uploaded on YouTube every minute (Statista, 2020). Huffff!!!!
2. A giant leap
Satellite imagery is not new. It has been there from around 1960. What has changed is the way we process those images. Earlier, the workflow was to download the images to your local machine and process it using GIS software. This was really challenging given the fact that high resolution satellite images are data heavy. Just imagine, how much space would be needed to store the images if you wanted to see analyze changes in precipitation, temperature or cropping patterns etc. over a decade. Downloading these images from the source and importing into GIS software is another time consuming task. Due to these logistical constraints, satellite image visualization and analysis never really took off until few years back when cloud computing was born.
Google Earth is a path breaking innovation in the area of satellite image visualization. It enables us to view satellite imagery, maps, terrain, 3D buildings etc by interacting with a virtual globe in 3D. Its timelapse feature allows us to experience the fourth dimension of time in which four decades of satellite imagery (24 million satellite photos) have been put together to showcase plant level changes (Source: https://medium.com/google-earth/time-flies-in-google-earths-biggest-update-in-years-183f3ea4ccf6).
A related tool is the Google Earth Engine which makes geospatial analytics lightning fast. All the satellite images are optimized for the cloud. Image processing which used to happen on local machine can now be performed on massive servers which deploy parallel computing to deliver blazing fast performance. The satellite data loaded on GEE like platforms has already been processed so it is ready for use in a wide variety of applications. This will open doors for more “big data” applications whereby vast amounts of satellite data over many years can be analyzed to look for subtle patterns that could not have been identified otherwise. The role of machine learning and artificial intelligence will be key in unlocking insights from such vast data stores.
3. Illustrative use cases
GEE catalogue is a one stop destination for years of satellite imagery and vector files. I will list some of the popular datasets here to motivate the possible applications of this platform. This is not anywhere close to exhaustive but is intended to ignite ideas where geospatial analytics can be employed for the greater good.
4. Conclusion
Any discussion on geospatial analytics will remain incomplete without Google Earth Engine (GEE). It is a tool which helps you visualize and analyze satellite images almost in real-time doing away with the inconvenience of downloading data from various space agencies. It provides programmatic access to petabytes of catalogued satellite images which can be filtered easily using metadata. It also allows to use own your own shapefiles for a more customized analysis. Handling satellite images has never been so easy. GEE enables us to perform computations on these images on the fly.
The native programming language for GEE is Javascript, however, Python lovers can also work with GEE by minor tweaking of the Java codes notwithstanding the scarcity of Python tutorials /code snippets in this domain. Please check out the code editor of GEE and its documentation to explore the power of GEE. For sample codes in Python illustrating possible use cases, you can check out my blog pieces published on my medium page.
Disclaimer: Views expressed in this blog are personal.