In the era of Big Data, the Chinese Academy of Sciences has set up the Digital Earth Science Platform (CAS-DESP) by pooling its data and computing resources to aid scientists and decision makers in organizing, processing and analyzing big Earth data. CAS-DESP includes three sub-systems: digital earth foundation platform, information integration and services sub-system, global geo-spatial data production sub-system.
Digital Earth foundation platform is the core set of components of the CAS-DESP capable of exploring massive amounts of geospatial data at global scale. Based on its high-performance 3D visualization engine, it provides the interactive ability to visualize complex scientific data, and real-time simulation of natural phenomena and process of dynamic systems including oceans, atmosphere, and biosphere.
Information integration and services sub-system provides the big data engine and service engine, which are integrated with the state-of-the-art distributive computing components for big data, such as Hadoop HDFS, Redis, and, Dockers etc. This sub-system also provides several efficient big Earth data management functions. Additionally, for efficient utilization of the resources, resource monitoring and value evaluation services have also been integrated in this sub-system.
Integrated with high-performance algorithms including deep learning algorithms and machine learning models, the global data production sub-system also provides land use, vegetation, water and atmosphere products with 7 spatial resolution (from 2 m to 5000 m) and 4 temporal duration (day, month, season and year) at a global scale. It enables fast and intelligent information extraction on remote sensing image archives at a global scale. These global products are the fundamental data to support the decision-making using CAS-DESP.
The CAS-DESP has played important roles in many scientific and decision support aspects. A global wheat pests and diseases monitoring report has recently been released based on the analysis utilizing the CAS-DSP platform. The report highlighted the occurrence and development of typical wheat pests and diseases in many wheat production countries. Using the deep learning network embedded in the CAS-DESP, the national steel plants and other industrial factories information extraction was automated using GF-2 high resolution remote sensing imagery which was challenging using traditional methods information extraction methods.
More efficient algorithms and models will be integrated into the CAS-DESP in the next 2 years, which will make the platform more intelligent. Applications of the platform in the fields, such as global change, natural disasters evaluation, and environment monitoring will be developed to support adopting sustainable practices in the future.
Figure2 Digital Earth foundation platform v1.0
Figure3 Desert Locust Monitoring Systembased on CAS.DESP
Figure 4 CASEarth Exhibition System