Application Of Decomposition Multi-objective Algorithm In Infrared Detection Of Spacecraft Damage

The hypervelocity impact of space debris threatens the safety of spacecraft operation in space.Therefore,it is very important to detect and research the impact damage of spacecraft.Infrared non-destructive detection technology has been widely used in the aerospace field.This paper will study and apply multi-objective evolutionary optimization of infrared detection of spacecraft damage.The research basis of this thesis is to simulate the hypervelocity speed impact of space debris on the spacecraft through ground impact simulation experiments,so as to perform infrared detection processing and analysis on typical spacecraft materials and typical Whipple protective structures after impact,and provide reliable ground research for spacecraft impact damage research.On this basis,the research focus of this thesis is on the application of multi-objective optimization theory in the spacecraft damage infrared feature extraction and feature extraction infrared image segmentation.The main research content is:First,in the infrared feature extraction of spacecraft damage,a multi-objective optimization theory is introduced.By considering multiple performance considerations between the transient thermal responses of temperature points in the category area,a multi-objective optimization problem of typical transient thermal response selection for infrared detection of spacecraft damage is constructed.This method solves the problem of insufficient representativeness of typical transient thermal response in the fast infrared detection feature extraction method,provides a more representative regional typical transient thermal response for feature extraction,and fully mines the damaged area from the infrared image sequence information.Then,based on the decomposition idea,the typical transient thermal response selection multi-objective optimization problem of the spacecraft damage detection is decomposed into several sub-problems.In the improvement scheme,the idea of opposite learning is used to co-evolve multiple sub-problems and solve typical transient thermal responses.This solution provides a more versatile spacecraft damage detection feature extraction method.Finally,in the follow-up processing of feature extraction infrared images,the paper carried out research on infrared image segmentation of feature extraction for theproblems of infrared image noise interference,blurred damage edges and weak damage details.At the same time,an infrared image segmentation model for impact damage with two-layer multi-objective optimization is proposed,so that the algorithm can segment the spacecraft’s damage area from the material background area as much as possible.It solves the problem of low efficiency caused by continuously debugging the weight vector,and lays the foundation for future research on quantification of damage.