Dim Target Extraction Technique In Dense Stellar Background

With the exploration and development of space,space has become an increasingly important resource.On the one hand,space provides a new platform for human scien-tific research,national defense construction and information communication.On the other hand,with the increase of human space activities,the existence of a large number of space debris seriously endangers the safety of spacecraft.Therefore,it is particularly important to monitor space debris and master the orbit and size information of space debris to ensure the safety of spacecraft operation.Space target detection is to detect the target from the complex stars background and obtain the target orbit and brightness information.In this thesis,the space target detection technology is studied.In this paper,ground-based photoelectric telescope is used to detect space debris,and the following problems exist,a)Ground-based telescopes are easily disturbed by skylight conditions and cloud,resulting in uneven background of frames.b)The un-evenness of CCD will cause the unevenness of image.c)The stars observed by the telescope will be affected by atmospheric dispersion and causes a vertical dispersion,d)The existence of dense stars in the field of view affects the efficiency and accuracy of target detection.e)The space target has low SNR and is easily drowned by noise.This brings great challenges to the extraction of space objects.Therefore,how to quickly and accurately extract dim spatial targets from complex background images is the focus of this thesis.This paper starts with the working mode of the telescope,analysis the imaging characteristics of space obj ects and stars in the frames,and combines the parameters of the telescope to complete the preprocessing technology of space objects,multi-target rapid extraction technology,extremely low SNR target extraction technology and target centroid extraction technology.The details are as follows:1.The motion and shape characteristics of background stars and debris under dif-ferent tracking modes are analyzed.The difficulties of space debris extraction are sum-marized.2.This paper analyzes the causes of image clutter,discusses the limitations of tra-ditional background suppression techniques,and proposes a background suppression algorithm based on local processing.This method can effectively remove the back-ground inhomogeneity caused by different skylight conditions,cloud background and CCD response,and the dark and underlined targets in the image can be completely re-tained,and good suppression effect has been achieved in the experiment.3.This paper analyzes the cause and influence of atmospheric dispersion,dis-cusses the correction method of atmospheric dispersion,and puts forward the atmo-spheric dispersion correction algorithm based on image deconvolution,which can elim-inate the influence of atmospheric dispersion and reduce the manufacturing difficulty of optical system.4.In this paper,a fast multi-target detection technique in dense stellar background is proposed.The multi-frame mask algorithm based on stellar registration is used to filter the stellar background,and the multi-frame track correlation is used to quickly extract the target based on the motion characteristics of the target.The effectiveness of the algorithm is verified by simulation and experiment.5.A new algorithm based on morphological expansion was proposed to extract weak targets and improve the detection effect of the target.6.This paper discusses the adaptability and limitation of traditional space obj ect centroid extraction algorithm,proposes multi-frame line fitting method which is suit-able for occlusion obj ect centroid estimation and contour ellipse fitting method which is suitable for adhesion obj ect,and verifies the effectiveness of the algorithm through experiments.