Star Image Simulation And Evaluation Of Star Point Extraction Algorithm For Star Sensor

With the rapid development of satellite application technology,the star sensor has been widely used in determining the attitude of spacecraft.Due to the limitation of development cycle and experiment cost,it is urgent to carry out the research of key technologies such as star map simulation and star point target extraction algorithm evaluation during the design of star sensor system,index demonstration and stereotypes test.In this paper,for the application of star sensor in orbit engineering requirements,based on the analysis of star target and stray light characteristics,the map simulation and implementation methods,star point target extraction algorithm performance evaluation research are carried out,the research work includes the following several aspects:(1)Analysis of stellar target and background characteristic.Based on the theory of basic radiation,stellar target and the transmission characteristics of background stray light are deeply studied.According to the direction of optical axis of the star sensor and the field of view,the visible star point is retrieved in the SAO star catalogue,which is then transformed by coordinate and magnitude.The gray scale diffusion of the target star point is carried out by the two-dimensional Gaussian distribution model,and the star point target simulation is completed by using the defocusing technique.The hybrid light suppression structural model is built in Solidworks,then import the model into Tracepro software,the parameters of the source is set.The simulation of the hybrid characteristics in different incident angle is obtained through the ray tracing,and the results provide input data for subsequent star map simulations.(2)Star map simulation of star sensor.Through the selection of detection device and the design of optical system,the star sensor system is designed.Based on the design parameters of star sensor system,the simulation results of stay light in different angles and the simulation results of star point target are superimposed to complete the star map simulation.At the same time,according to the orbital parameters of the satellite,the orbit parameters of the sun and the observation time,calculate the angle of the camera's optical axis direction and incident angle of the sun,superimposed star map simulation image and the stray light simulation image in corresponding angle.The dynamic star map simulation of star sensor is achieved.Finally,the star map simulation software is designed.(3)Implementation and evaluation of star point target extraction algorithm.A series of image preprocessing operations on simulated star maps are achieved by gray scale statistics,threshold segmentation,connected domain partitioning,and effective separation between star point target and sky background,single star target and other targets are completed.Using the centroid algorithm,the improved centroid algorithm and the Gaussian fitting algorithm,the three kinds of star point extraction methods are used to locate the centroid of the star map and calculate the centroid coordinates of the star.The simulated star map under the influence of noise and stay light is analyzed,and the accuracy of the target extraction of the three algorithms are evaluated.In summary,through the stellar and background characteristics analysis,star sensor star map simulation and star point target extraction algorithm performance evaluation and other series of research work,the star sensor in orbit work process are simulated,and the star target extraction algorithm evaluation is achieved,which provides the theoretical basis and technical method for the design of star sensor system,and the testing and evaluation means for the factory identification of star sensor products.