Research On The Key Technology Of Space Debris Monitoring Based On Vision

With the rapid development of aerospace technology,the exploration of outer space environment is increasingly frequent,and the space environment has become a forefront of competition among countries.In order to ensure the normal operation of the spacecraft,it is necessary to strengthen the perception and monitoring of the space environment.,and the vision-based space debris monitoring system must be an essential measure.This paper carries out corresponding research on the identification and tracking of debris in the spatial background,focusing on the preprocessing of complex spatial images,the detection and recognition of debris in the moving background and other key technologies.First,in order to remove the complex poisson noise in spatial images,considering that Markov Fields of Experts can accurately describe the prior probability of natural images,and then the Markov Fields of Experts model is used as the prior term.A poisson denoising model based on the Bayesian maxi-mum a posteriori model is established.Compared with the traditional method,the proposed algorithm improves the signal-to-noise ratio by at least 0.18 d B.In order to eliminate the stray light in the spatial background,an image stray light elimination algorithm based on expectation maximization was proposed based on classification and segmentation.The optimal parameters of the gaussian mixture model were re-transformed by the EM algorithm to realize image labeling and separate the targets in stray light.Through comparative experimental verification,the method proposed in this paper can effectively detect the target in the stray light environment with a very low rate of missed detection.Secondly,due to the shortcomings of modeling the fragments in the center as the background by the gaussian mixture model,an improved gaussian mixture model fragment detection algorithm based on wave gate tracking is proposed.In the gate region,the central target is detected based on the gray level.Through experimental verification,the improved method in this paper improves the shortcomings of traditional gaussian mixture model,and can achieve real-time detection of central targets.Since there are debris and stars in the detected objects,their geometric features are very similar,in order to identify the debris targets,the motion differences between the stars and debris are analyzed,and a debris recognition algorithm based on the anomalous characteristics is proposed.The test results show that the proposed algorithm can effectively identify debris and stars.Finally,combined with the fragment imaging model,the weight centroid method is used to locate the fragment centroid coordinates.Aiming at the scene that produced the smear effect,a method of extracting the centroid coordinates of the fragment body after cross-projection is proposed,which improves the accuracy of the fragment centroid location.In order to obtain the trajectory of debris,a multi-target tracking algorithm based on cost matrix and Kalman filter is used for multi-target trajectory correlation.The working mode of the space debris monitoring system is designed,and the system software module is built and verified by simulation.Experimental results show that the system designed in this paper can realize real-time detection and tracking of space debris.