Research On Terahertz Image Enhancement And Fusion Algorithms For Internal Defects Of Aircraft Skins

Internal defects in aircraft skin are one of the important factors affecting the safety of civil aviation operations.Therefore,improving the quality and efficiency of internal defect detection in aircraft skin is an urgent problem to be solved.Terahertz imaging technology has unique advantages such as non-contact and strong penetrability,which can achieve non-destructive testing of internal defects in aircraft skin.However,terahertz imaging technology still has limitations in detecting and characterizing more complex defects at deeper levels.Therefore,it is necessary to enhance and fuse terahertz images to achieve automatic defect recognition and quantitative evaluation.Based on this,this thesis uses image enhancement algorithms and image fusion algorithms to process terahertz images and conducts a series of related theoretical analysis,algorithm research,experimental simulation and other work.Specifically,it includes the following points.(1)This passage describes a proposed method for enhancing THz(terahertz)images that addresses issues such as low image resolution and high noise levels.The method combines a local contrast algorithm based on a central surround framework with a bilateral filtering algorithm.The first step focuses on improving image contrast and clarity by using the local contrast algorithm based on the central surround framework to enhance the THz image.Next,the bilateral filtering algorithm is used to reduce noise in the THz image while retaining the edge details.Finally,experimental comparisons are conducted to evaluate the ability of the proposed image enhancement algorithm to improve THz image contrast and enhance the edge details of defects.(2)An improved adaptive image fusion algorithm is proposed to address issues such as texture loss and blur degradation in THz images.Firstly,preprocessing is performed on the THz defect image,and the source image and the preprocessed image are used as the input images for the image fusion algorithm.The fusion image is estimated based on the relative difference in local contrast between the two THz images.Then,a Gaussian filter is used to extract non-spectral spatial details from the preprocessed THz image,while the Sobel operator is used to enhance the detail information of the image,making the spatial details more prominent.Finally,the extracted details are weighted according to the fusion image and fused into the THz image to obtain the fused THz image.Experimental results show that the proposed image fusion algorithm suppresses noise and improves the quality of THz images.