Design Of Real-time Preprocessing System For Meteorological Satellit

In order to realize the efficient application of meteorological satellite observation data,lay the foundation for the subsequent on-site remote sensing information extraction and information generation technology,and carry out the geometric positioning and radiation calibration of the uncontrolled point in real time for the payload observation data,this paper selects the main optical load of the meteorological satellite,the medium-resolution spectral imager,as the processing object,and constructs an on-board real-time preprocessing system based on space VPX architecture.The system selects FPGA+DSP as the core processing module to complete the design of the entire process from data input,data processing and data output,and finally realizes the real-time output function of first-level positioning and calibration product data.In order to improve the processing rate of the system,the application capability of the system is extended to multi-load and multi-channel scenarios,and the inter-core communication mode of multi-core DSP is optimized,which solves the problems of data consistency between cores and waste of shared storage resources,shortens the processing time of the system and improves the memory utilization.The main work of this paper is as follows:1.In order to improve the timeliness of satellite applications,shorten the time of payload data transmission and ensure data validity have become the key ways to develop spaceborne processing technology.2.Designed a spaceborne real-time preprocessing system based on Space VPX architecture,using FPGA+DSP as the data processing board architecture,completed the main chip selection and hardware scheme design,and divided the system software part,of which DSP is responsible for realizing the positioning of load data,calibration core algorithm,FPGA is responsible for realizing data scheduling interface,and the SRIO,EMIF,RS422 and other interfaces are designed in detail.Complete the design and simulation verification of geometric positioning and radiation calibration algorithm for the observation data of FY-3D medium resolution spectral imager,and realize source code development based on hardware platform3.Aiming at the characteristics of the MERSI imager with multiple channels of data and the continuous output of the data stream,the pre-processing system in this paper uses the advantages of DSP multi-core parallel collaboration and FPGA parallel processing to realize DSP multi-core parallel processing by data flow drive mode,and completes inter-core communication by means of shared storage query.Considering the scalability of multi-load data processing,a DSP inter-core communication scheme based on the combination of circular queue structure and shared storage query mechanism is designed,which solves the problem of data consistency,and designs a multi-core DSP shared storage partition mode,which optimizes the processing performance of the on-board data preprocessing system,and doubles the processing rate compared with the traditional solution.4.The on-board real-time pretreatment system test platform is designed,which realizes the input of load and satellite platform data,the processing of the pretreatment system,and the output process of the result,and obtains the processing results that are consistent with the simulation results of the algorithm,and the 1-cycle positioning and calibration processing time of the system is 954.5ms,which meets the generation index of 1.5s single-cycle data.The correctness and real-time accuracy of the on-board real-time preprocessing system are verified.