Research On Image Processing Algorithms Of Orbit Determination Of Synchronous Satellites Based On Drift-scan

When the synchronous satellites are being observed,to overcome the disadvantage that background stars and synchronous satellite can not be achieved round shape at the same time,short frame CCD with drift-scan model is proposed,using the starring model and drift-scan model alternately observing background stars and synchronous satellite to present round shape in star images,then processing star images.To realize real-time orbit determination of synchronous satellites,DSP image processing module is introduced to process the star images rapidly.On the platform of DSP,the image processing software is developed.The main procedures of image processing include 4 parts: image preprocessing,targets extraction,star image identification and orbit determination.In this paper,the targets extraction and star image identification is described in detail to reach the demand of high-precision and real-time orbit determination.The main research contents include: firstly,eliminating the un-uniformity brightness of image background by using the method of polynomial fitting,and finding and elimination the smear trails by comparing mean column gray value;secondly extracting the small targets based on visual saliency algorithm,furthermore,optimizing the algorithm to accelerate the extraction speed and improve the extraction precision;thirdly matching the star targets with reference stars in star catalogue by using twice match method to increase match rate as well as decrease match consuming time;finally,analyzing the orbit precision after calculating the orbit based match outcomes.By processing the star images taken by drift-scan CCD,the experiments results show that the algorithms proposed in this paper can process the image efficiently,extraction targets and identification images quickly,moreover,the extraction precision and identification rate are high.The final orbit determination outcomes have high determination precision,which can meet the challenge and validate the algorithms.