Research On The Attitude Accuracy Improvement Of High Resolution Pushbroom Satellite

With the rapid development of technology,remote sensing satellite images are moving towards higher spatial resolution,faster time resolution and higher geometric accuracy.China’s remote sensing industry has been brought to the level of centimeter resolution by the ZiYuan-3 satellite series,TianHui satellite series and GaoFen satellite series,and the accuracy of geometric positioning has reached 10 meters.However,the further geometric accuracy is greatly challenged,and the accuracy of the attitude processing of the existing satellites has already restricted the geometrical positioning accuracy.Through the in-orbit satellite image data,this article tries to deal with the problem of satellite attitude processing,and to analyse and classify the attitude errors theoretically.Then it explains the concrete phenomenon of problems,and the solving ways.The main research contents are as follows:(1)The low and high frequency errors in attitude processing are analyzed.The traditional attitude processing research mainly focuses on the attitude determination algorithm,and the accuracy and stability of the attitude are verified by simulation.In this thesis,the low frequency and high frequency error of satellite attitude processing are put forward from the perspective of satellite image geometric positioning accuracy.Firstly,aimed at the Angle changing between the star sensor optical axis and the camera’s optical axis,the low frequency error both from star sensor itself and from the stability of the camera structure is proposed.Then the interpolation error caused by satellite attitude stability and the attitude error caused by satellite platform jitter are called high frequency error.The influence of attitude error on geometric positioning accuracy is analyzed through the 3-axis of satellite coordination.(2)A long time/wide-area control method is proposed to verify the low-frequency attitude error of the in-orbit satellite.The precision of the long time monitoring results show that the ZiYuan-3 image accuracy decrease as far from the calibration time.Based on the theory of external calibration method,the thesis adopted different calibration parameters in different periods,and effectively improve the positioning accuracy without control through a long time monitering of the images.The thesis globally selected 11 track data of the ZiYuan-3 strip images,through the control data and high resolution images at home and abroad,to verify their geometric accuracy.and established the position precision distribution curve based on latitude,which find that it has many order of Fourier fitting law of low-frequency,and the law is more obvious in the southern hemisphere.(3)This thesis proposes a method to improve the relative attitude accuracy based on small matrix CCD,which can realize effective compensation for low frequency error.In this thesis,the installation mode of load with a small matrix CCD is presented,which can be observed the small matrix CCD images and linear array images simultaneously.Based on the matching of adjacent images,the relative attitude of high precision is restored.Under the support of the theory of stereoscopic intersection,and the relative attitude recover from the small matrix CCD,the thesis concluded that the symmetric stereoscopic has a highest precision in the case of the same orbit.The experiment analysis the influence between the symmetry of stereo pairs,base height ratio,attitude stability,and the sampling rate with the intersection accuracy.Through the experiment of GaoFen-9 stereoscopic intersection accuracy under different staring angle,and the experiment of after the relative attitude precision increases under small matrix CCD,the vertical accuracy increases under the symmetric stereoscopic,the thesis verify the correctness of the theory.(4)A method of high frequency attitude jitter elimination is proposed.This thesis starts from the analysis of the jitter source of satellite platform and discusses the influence of 3-axis jitter on image.Then,the method of detecting high frequency jitter by different-time imaging data is established,and the high frequency correction method based on accelerometer measurement data is proposed.In the experimental part,the validity of the method is verified through the simulation and processing of different high frequency combination,and then the satellite images with high frequency jitter are detected by the in-orbit satellite data.Through the measurement data of the accelerometer of the GaoFen-9,the influence of the jitter on the image was corrected.Finally,the validity of high frequency jitter correction is verified by means of road straightness detection and different-time imaging data matching residuals,which indicate the effectiveness of the method of high frequency attitude jitter correction.