Research On High Precision Geometric Processing Methods And Application Of Spaceborne SAR

Because of its all day,all weather and global observation capabilities,the space borne SAR with satellites and other space vehicles as the motion platform has become an indispensable means of space to earth observation.This technology has played a very important role in the fields of cultural heritage protection,marine monitoring,military reconnaissance,geological mapping,agricultural and Forestry exploration,environment and disaster monitoring.Since the first satellite SAR satellite(Sea Sat)was successfully launched by the United States in 1978,many countries around the world have developed and successfully launched several SAR satellites.With the emergence of new modes,new systems and new technologies,space borne SAR has the characteristics of wide range,high resolution and multi-dimensional,which promotes the vigorous development of space borne SAR.Our country starts late in the field of space borne synthetic aperture radar(SAR)technology,and the data of space borne SAR image can not be self-sufficient for a long time.In recent years,with the launch of the major special project of high-resolution earth observation system in China,the number of SAR satellites in orbit is increasing.The successful launch of gaofen-3 in2016 has brought China’s space-based remote sensing into an all-weather,quantitative and meter level application era,greatly improving the supply of domestic SAR Remote Sensing data.However,on the one hand,the performance of domestic satellite SAR devices is weaker than that of foreign countries.On the other hand,Satellite SAR imaging processing and geometric processing belong to two different fields,so it is difficult to achieve effective collaborative processing.As a result,the geometric processing accuracy of domestic SAR satellite images is not enough,which has been criticized for its poor geometric quality for a long time.The key factor that determines whether the application efficiency of SAR images can be fully exerted is the geometric positioning accuracy.Therefore,it is urgent to establish a perfect system of highprecision geometric processing methods for Spaceborne SAR.The purpose of this dissertation is to achieve the decimeter geometric positioning accuracy of spaceborne SAR(especially for domestic SAR),and to establish a complete set of high-precision geometric positioning processing methods for Spaceborne SAR,including the fine modeling of end-to-end error,the introduction of error modeling and compensation for geometric positioning considering the stop and go approximation of imaging processing,geometric calibration based on geometric calibration field and geometric calibration without geometric calibration field support Research on the processing of standard and stereo SAR,etc1)Study on the characteristics of all link geometric positioning error sources of spaceborne SARThis dissertation studies all kinds of error sources in the process of Satellite SAR earth observation,analyzes the influence of non ideal factors such as SAR sensor performance,satellite platform orbit positioning,echo signal atmospheric propagation delay and observation target coordinate offset caused by geophysical effect on geometric positioning accuracy,and models the end-to-end error characteristics.In view of the insufficient connection between spaceborne SAR imaging processing and geometric positioning processing,which leads to the lack of geometric positioning error processing,this dissertation analyzes the positioning error caused by the stop approximation effect in imaging processing,and puts forward the corresponding compensation scheme.Finally,the experimental verification is carried out by sentinel-1 and yg-sar data.2)Research on error compensation model of geometric positioning system for Spaceborne SARBased on the geometrical positioning error characteristics of spaceborne SAR,the system error compensation model is constructed,and the solution method based on the on-board calibration loop and the geometric calibration field is studied.According to the SAR signal theory,an accurate coordinate extraction method of the angle reflector on the SAR image is proposed.In view of the limited hardware performance of domestic spaceborne SAR,a multi time series joint geometric calibration parameter solution method is proposed.Through the real SAR image experiment,it is verified that this method can effectively solve the problem that the calibration parameter accuracy is not ideal due to the random difference of sampling delay on the satellite.3)Research on geometric calibration method of spaceborne SAR without geometric calibration field supportAiming at the problem that the traditional geometric calibration relies heavily on the ground high-precision control,and the increasingly urgent need of normalization calibration,two geometric calibration methods of spaceborne SAR without the support of ground geometric calibration field are studied.Based on the error reflexivity,a self calibration method with symmetric geometry constraints is proposed.A self calibration method combining non photogrammetric constraints is proposed by using some more general original non photogrammetric observations or conditions in SAR images.Through yg-sar image experiment,it is verified that the accuracy of field-free geometric calibration is slightly lower than that of field-based geometric calibration.4)Global high precision robust control point(GCP)generation using space borne SARThis dissertation introduces the processing flow of conventional stereo SAR,aiming at the fact that the high-precision positioning of conventional stereo SAR depends on the external meteorological observation value and precise orbit data.In order to improve the real-time performance,a differential stereo positioning method is proposed,which realizes the high-precision stereo positioning on the basis of a reference point.To study the feasibility of all kinds of PS points existing in urban area as SAR candidate robust GCP;because it is difficult to recognize the same name PS points based on SAR data only,a method of automatic recognition and extraction of the same name PS points of stereo SAR using optical satellite image is studied,and error points are eliminated based on phase noise estimation.