Research On Spatial Point Target Detection Method In The Large Field Of View

This thesis focuses on the research of spatial point target detection technologyin the large field of view. In this thesis, the characteristics of stars background andstar small targets are deeply analyzed; the noise performance of specific detector issystematically modeled and analyzed; the various characteristics of star small targets(stars and space point target) are also analyzed, including the imaging property ofstar, stellar spectral characteristic, the motion characteristics of the star, and imagingproperty of space point target, etc. Meanwhile, cross projection and thresholdcentroid methods are adopted to analyze the star map processing algorithmperformance. Star positioning accuracy will also be analyzed.For spatial point target characteristics in different orbital positions, three spacepoint target detection algorithms applicable to large field of view are proposed inthis paper, which are motion space point target detection method, space point targetdetection method in high orbit, and space target detection based on star patternrecognition. In the first method, single-frame and multi-frame star map targetdetection methods are given to be aimed at the characteristics of the motion (in LEO)space target. In the second method, the reference map is simulated through theknown probe optical axis; and then the stars registration is used for star centroiddifference to the reference map and observation map, and candidate target points areextracted; finally, the real target points are determined by the continuous star map“choose3from5” principle. In the third method, part of the star points are identifiedby the star pattern recognition applicable to the large field of view; and then thedetector attitude is calculated by QUEST method; besides, by means of this opticalaxis, the real target points are extracted and detected by adopting the methodproposed in section4.2.The simulation experiment and analysis on the performance of these threespatial target detection methods by MATLAB show that detection methods proposed in this thesis can be well applied to detect the space point target under the large fieldof view.