Research On Infrared Closely Spaced Objects Super-resolution And Location

Remote imaging system for space target imaging,because the observation distance,the image resolution is low,if the target space distance from each other closer,,and the Closely spaced object(CSO)is formed due to the observation distance and the target space distance is close to each other,which makes it difficult to obtain the information of the number and location of the target.Closely spaced object super resolution technology can divde close objects in the image and contribute to the separation and localization of the target.In this paper,the super resolution tracking technique of adjacent targets in infrared space is studied,and the imaging process of space adjacent targets is simulated.An improved method of super-resolution of adjacent targets is proposed.The main work of this paper includes three aspects: 1)Modeling the imaging process of adjacent target in infrared space,and establish an infrared image simulation and positioning system.The system realizes the satellite simulation of space moving target imaging,which includes the relative motion of satellite and target,3D target project to plane process,imaging model simulation,noise simulation?The system can simulate infrared images similar to real scenes.2)studied the infrared CSO super-resolution and positioning technology,proposed a method pre-process the image then super-resolution.Including image denoising,target area extraction,CSO super-resolution,etc.,and from the positioning error and time-consuming aspects of the performance of the algorithm were evaluated.3)The CSO tracking technique is studied,and a simple tracking method is proposed to estimate the target motion.The target trajectory can be estimated with fewer frames,and the tracking of the target is continued.Finally,the performance of the tracking algorithm is evaluated.The infrared CSO simulation and super-resolution tracking system realized in this paper are of great significance to the design of the super-resolution algorithm which is accurate.It can make the algorithm designers more easily evaluate the performance of the algorithm and help the improved algorithm.