The ground processing system of the high-resolution satellites is an important component of the high-resolution earth observation system. Following the principles of standardization, safety and reliability, maintainability, economy, and advanced technology, a remote sensing data comprehensive processing platform, which can realize the functions of data ingestion, multilevel standard data production, archiving, simulation, evaluation, distribution services, radiometric and geometric calibration, observation management of the payloads on GF-1 to GF-7 satellites, is designed and developed.The ground processing system adopts the cloud computing ground processing architecture based on the HADOOP platform and integrates the internal storage of nodes into a massive data storage system through the HDFS distributed file system. It also realizes the united scheduling and large-scale parallel processing of massive data through the Yarn resource management system. Moreover, it studies and develops the high-precision processing algorithms of the multispectral, hyper-spectral, infrared, microwave, Lidar, and atmosphere detection payloads on the satellites.The high-resolution satellite ground processing system builds a multisatellite comprehensive processing platform for massive satellite data. The capabilities of the ground processing system have been substantially improved in data integrated processing, data storage and management, and data distribution services; it provides standard data products to users with high accuracy and quality.It has 400 trillion calculations per second and highly automated data processing capacity of more than 20 TB per day. It can process a wide range of satellite payload types for GF-1 to GF-7 satellites. In the normal mode, the data processing time of each orbit is not more than 2 hours, and the emergency processing time is not more than 45 minutes.The system has 25 PB scalable storage and management capabilities of massive remote sensing image data. The products are stored online and near line within the life cycle, and level 0 data are permanently archived.For data distribution, the system can distribute more than four million scene data every year and supports 1000 external users to visit and retrieve the data on the service website at the same time. The average retrieval response time is less than 10 seconds. It also supports 200 external users to download data products online.The system has the continuous operation ability of 7×24 hours. The operation stability is better than 99%. The main and backup switches of the key business equipment are completed within 10 minutes after the failure. The mean time to repair is 4 hours.The high-resolution satellite ground processing system has realized satellite data ingestion, standard product processing, archiving, simulation, evaluation, distribution, calibration, payload management, and data sharing services. It has also realized the comprehensive data processing of a wide range of remote sensing data, which cover optical, submeter high spatial resolution, multimode radar, geosynchronous orbit, hyperspectral resolution, atmospheric, stereo mapping, and Lidar earth observation. A multisatellite and multiload remote sensing data product system have been established to provide multisource remote sensing data for land, ocean, agriculture, disaster reduction, forestry, water conservancy, environment, earthquake, meteorology, transportation, surveying, and mapping users. It has played an important role in promoting China’s economic construction, social development, scientific and technological progress, and serving China’s belt and road initiatives.

cloud computing;high-resolution satellite;ground processing system;remote sensing data processing