Research On Aerospace Composites Defect Detection Technique Based On Multi-feature Weight Vector Space Model

Due to the infulence of Composites,Manufacturing Process and Working environment,aerospace composite structural components are prone to produce different types of defects,which ultimately causes aviation equipment failure and even leads to accidents.In order to select the effective prevention and treatment strategies timely,prolong the life of aviation equipment,reduce the accident rate,it is essential to carry on defect detection for aviation structural components.Advanced defect detection technology could timely identify deficiencies,thereby ensure the quality and safety of aerospace structural components.Thus,in recent years, defect detection technology is widely applied in the aviation body, steel, wood, textile, traffic and other fields.The main contents of this paper include image enhancement,edge detection,clustering segmentation and defect classification in the defect detection technique.On the aspect of image enhancement,use Bilateral diffusion model of Partial Differential Equations to smooth non-edge portion and sharp edge portion to enchance the image.On the aspect of edge detection,apply quantum superposition model that incorporates the edge feature for edge detection to ensure the integrity of edge details.On the aspect of clustering segmentation and defect classification,first,use fuzzy clustering segmentation algorithm to extract defect target,then classify the defect image into different types according to a combination of geometric features and defect type definition.Innovative points of this paper include the following aspects:(1)Due to the impact of unreasonable Shock term in the anisotropic diffusion filter model, it is prone to produce spots, ring effects and other issues when enchance the image. Therefore, a bidirectional diffusion model merged with local edge feature is proposed.The model is based on the edge feature selection strategy, which overcomes the deficiency of the two order zero crossing point as the selection diffusion strategy.The experimental results show that the proposed model can effectively enhance the contrast of the image.(2)The traditional edge detection algorithm is sensitive to noise and its accuracy of edge detection is low.Thus,a quantum mechanical edge detection model based on the fusion edge regularity is carried out.The model is mainly based on the three quantum bit space,which is constructed by the edge tangential continuity, the normal direction discontinuity and the difference between target and background.Experimental results show that compared with the traditional edge detection algorithm,the quantummechanical edge detection model could detect edge more accurately.(3)In view of the shortcomings of traditional defect classification algorithms, such as artificial setting parameters, poor anti-noise ability and low classification accuracy, a defect classification model based on FLICM and geometric features is put forward.The model takes advantage of the FLICM’s high segmentation accuracy and geometry features’ s flexible combination to classify the defects.The results show that the model can achieve the correct classification results of all kinds of defects such as inclusion,porosity, crack and so on.