Big data: No, how to use it?

The satellite collects so much data, if not, what do you collect it for? If only a handful of people closed the door, what is this secrecy supposed to prevent?

Large data is used to protect the environment, and researchers develop new forest mapping tools that warn the monitors when large tracts of forest are felled and destroyed.

Now the domestic big data is called very loudly, but the empty talk is more. I don't like empty talk, I want to see application examples of large data.

The Spectrum alert, April 16, published an article on the use of large data for environmental protection, and researchers developed a new forest mapping tool that warns the monitors when large tracts of forest are felled and destroyed. This aroused my interest.

This online tool, developed by the World Resources Institute, called Global Dara Watch, which uses large data-processing techniques for satellite imagery, can detect deforestation and damage at any time. Through it, users can explore the global map, discover the trend of forest change since 2000, and can study the forest situation, its precision can reach 30 meters. The tropical area of the map is updated every 16 days. Users can choose the "alert" feature, which will alert you when the system detects illegal logging or forest fires.

When a tree falls in the forest, the computer sends a warning to environmentalists, researchers and environmental policymakers. It seems to be easy, but it's hard. To find that particular forest from the messy, specially-formatted satellite data, and then identify the tree that just fell from the image, send a warning in real time to the person involved-a lot of technical difficulties.

Global Dara Watch provides technical support from the Google Maps engine, providing image data from several satellites of NASA and the United States Geological Survey. Previously, Google developed a platform to centralize the petajoules of data from Earth science and to open it to researchers so they could use it in a straightforward way. "Users just log in, take out the data, and run their own algorithms," says David Thau, a senior developer at Google Earth engine. "Google Earth engine now has thousands of research partners.

And Google's forest surveillance is an integration of a series of project research results. The data laboratory of the World Resources Institute has long had a hotspot area forest alert system, based on data measured by the medium resolution spectrometer of two satellites Terra and aqua. Professor Hansen, a geo-science Matthew at the University of Maryland, worked with Google Earth engine to use several Earth resources satellites from NASA and the USGS for global forest surveillance. The former can provide better time resolution, while the latter can provide a good spatial resolution.

2008, the United States Earth Resources satellite data are free to use. Every 30.1-meter pixels, a total of 143 billion pixels, Hansen tracked the dots every month, finding that their characteristics would change according to the seasons. Hansen's collaborators published a report in November 2013 on a 1.5 million-square-kilometer forest that disappeared from 2000 to 2012, calculating 1 million CPU hours on 10000 computers. In cloud computing, researchers have to consider the distribution of computing tasks across the network, and the researchers of the Earth Engine can automate parallelization by simply using a program interface to access their problems. In order to build a common global forest surveillance website, the World Resources Institute plans to make these big data available to the public, and it is hoped that government departments, businesses, researchers and various interest groups can use this tool to get better images of forest management.

Through this report, I have a few thoughts:

A large number of data can be applied here to determine the three elements: satellites, David Thau and Matthew Hansen. Satellite provides data, David Thau development platform, Matthew Hansen do data analysis.

Second, satellites are to be used. We often hear about the success of our satellite launch, but we rarely hear the news that satellites are being used. Our satellite launch technology has indeed been cleared and has been commercialized. We have also launched a number of satellites ourselves. However, little is heard of what satellites do, what they do to the national economy, and satellite data can be used freely by researchers.

Third, the study of large data must have practical use. Using large data for forest monitoring is a good example. Every year China's forest fires and illegal logging occurs, if automatic monitoring can be achieved, and real-time to the relevant personnel to push warning, that how good ah! To achieve this, there is much work to be done--it personnel to provide tools to facilitate professional processing of these data; professionals have to study their own algorithms to make various monitoring judgments. It's not easy!

(Author: Chinese Academy of Sciences, Institute of Computational Research)

(Responsible editor: Meng)