Google Earth Engine - Image Analysis for the Masses?  My Take

For a while now, the launch of Google Earth Engine (GEE) has been rumored and indeed yesterday the announcement was made on the Google blog. Developed in Google Labs, GEE takes the archives of satellite image data and selectively processes the data for specific applications such as forest density and climate change. In fact, the launch was done at the United Nations Climate Change Conference in Cancun, Mexico. According to the blog:

Google Earth Engine is a new technology platform that puts an unprecedented amount of satellite imagery and data—current and historical—online for the first time. It enables global-scale monitoring and measurement of changes in the earth’s environment. The platform will enable scientists to use our extensive computing infrastructure—the Google “cloud”—to analyze this imagery.

In its current form, users of GEE can view imagery that had been processed previously with some basic image classification. Landsat and MODIS imagery are available for viewing from many choices in a limited image catalog. Users can select imagery from the catalog and view the data in the GEE workspace. Many organizations were involved in the image processing effort and they should be commended for this project. It is truly a step in bringing image classification to the masses.

But while this effort is a tremendous step forward, most geospatial remote sensing scientists will have to wait for what we really want…real-time, online image processing. You can’t do much with the GEE platform right now. And that will only come with the release of the GEE API that is now only available to a small handful of partners. You can email them if you want access.

Right now, you’ll have to be satisfied with simply viewing the results of image processing by others. But I had a hard time discerning what the imagery was trying to portray. The lack of a legend for each classified image hinders interpretation. Sorry, but a map isn’t a map without a legend. You can choose an image to view, zoom and pan, but not much more. Viewers will have to go to each project website for more details (a link is supplied with some project imagery). Check out the Map Gallery for examples.

I think that with this first step you can envision a time that raw image data will be classified and shared with millions of people and it will be yet another advancement in the education of many more people about the value of satellite imagery…just as we found out with the launch of Google Earth. We’ll have millions of school aged kids doing image classification for 5th grade environmental science classes. It will make the two years I spent doing image classification for my master’s thesis look like…well, child’s play.

It’s a great step forward but it leaves me wanting more. I guess I want to be that kid who wants my own image classification sandbox and I’ll have to wait for GEE to give me that opportunity.