Astronomical Target Detection And Tracking Algorithm Based On Image Recognition

Since the beginning of the 21st century,the enthusiasm of human beings in the development and utilization of space resources has been continuously rising.With the influx of a large number of funds and talents,the outer space of the earth is gradually becoming a new place for the competition between science and technology and military.In this new situation,major countries are trying to develop space technology and build their own space industry.Space technology can not only reflect the strength of science and technology,economy and other aspects,but also is closely related to a country’s national defense security.Target monitoring is very important in the space technology research field of a branch,its main task is to detect,track,and space target recognition in earth orbit,to determine their type,serial number,posture and trajectory parameters,such as information,the technology can not only help us understand the earth’s outer space,but also in future space attack-defense occupies an important position.Traditional tracking and control schemes need to prepare a large amount of target-related data in advance,so they are limited in application scenarios.At the same time,due to the open-loop control in the control,it is easily affected by external factors and the cumulative effect of errors,and the late operation and maintenance costs are relatively high.In order to solve the above problems,a new idea is to use the popular image recognition technology to track the target.Through wide field of target search in the first place,when the target in telescope view after switching to the integrity of field observations,the control system through analysis of CCD to get images to get target real-time position and then adjust the telescope pointing and repeating the process,finally realizes the function of tracking control telescope.Because this method successfully transforms the original open-loop control system into closed-loop control system,and does not need to prepare a large amount of information of the observed target in advance,the tracking stability and flexibility are improved.With the continuous improvement of imaging device resolution and computer processing power,the spatial target tracking scheme based on image recognition is gradually being widely used,and will maintain a rapid growth trend in the future.This paper mainly studies the relevant operations in the space target monitoring system based on image recognition,and focuses on detection and tracking,and designs a space target monitoring algorithm that can be used under complex background conditions.The algorithm in this paper focuses on the two objectives of detection and tracking,and carries out systematic and in-depth research work,which fully reflects the uniqueness and integrity of research ideas and methods.The innovation of this paper is mainly reflected in the following two aspects:(1)Aiming at the complexity of starry sky background,an image time domain preprocessing algorithm based on ORB registration and difference is designed.Firstly,the moving background is compensated by registration,and then the moving target is extracted from the background by difference.According to the experiment of video processing of spatial moving target,the results show that the proposed algorithm can eliminate about 95%of the false target signals in the picture,so it can greatly reduce the computational pressure in the stage of target detection,and improve the real-time performance of the system.(2)For the design of tracking scheme for small space targets,this paper adopts the combination of KCF target tracking algorithm and Kalman tracking algorithm,in which the KCF tracking algorithm is responsible for high-speed tracking of the target under good conditions,and temporarily switches to Kalman filtering tracking when the target disappears.Thus,the failure of KCF tracking caused by the loss of small targets due to external interference is avoided.Experimental results show that both the traditional KCF algorithm and the proposed algorithm can track the target stably when the target moves along a straight line and no loss occurs,and the proposed algorithm has a higher degree of center coincidence.When the target is lost,the traditional KCF algorithm fails to track,but the algorithm in this paper can continue to track for a period of time,and the anti-jamming ability of the algorithm has been improved to a certain extent.