Research On Surface Defect Detection Of Underwater Structures Based On Image Processing Technology

Due to the complex environment in which underwater structures are located,they will be damaged to varying degrees during the service.Once they are lost,they will cause casualties and huge economic losses.Regular inspection of underwater structures is of great significance in ensuring the safety of underwater structures in service and giving full play to engineering benefits.With the rapid development of computer vision technology,people try to apply it to the detection of underwater structural diseases,but a unified detection method has not yet been formed.This article focuses on the problem of disease identification and disease area statistics of underwater structures.Based on computer image processing technology,an image processing technology combining dark channels and Gabor pyramids and a DR algorithm are used to expand by contrast.Research on the identification of diseases of underwater structures and the statistics of disease areas.The main research results of this paper are as follows:(1)Conduct theoretical and basic research on computer image processing technology.First explored the underwater image imaging model,and then introduced image enhancement and restoration from the spatial and frequency domains,introduced the three types of image segmentation from the threshold,region,and edge,and selected two underwater images as the basic images for effect display;(2)An image processing technology that merges channels and Gabor pyramids is used.Combining the advantages of the dark channel algorithm and the Gabor pyramid theory algorithm,the surface defect area statistics are performed on the processed image with obvious contrast in which the disease position becomes white and the color of the intact part becomes black.The surface defects of the 4 disease images selected in this paper accounted for 6.59%,2.65%,5.28%,and 5.73% of the total inspection area respectively.After actual operation verification,the image processing combining the dark channel and Gabor pyramid theory can achieve the purpose of detecting the surface defects of underwater structures;(3)A DR algorithm that can enhance the defect images of underwater structures is adopted.The DR algorithm enhances the details of the image,and combines the binary morphology theory and the optimal threshold segmentation theory to screen the edge of the disease to remove the interference information.After actual operation verification,it is concluded that the diseases in the 6 underwater structure pictures are 3.44%,1.68%,2.22%,4.59%,21.81% and 18.46%.The method can realize accurate statistics of the proportion of diseased pixels,and achieve the purpose of quantitative detection of surface diseases of underwater structures.Based on image processing technology,this paper realizes the accurate recognition of surface defect features.There are certain advantages in the image processing process,relatively high-quality target images can be obtained,which can solve the problem of dangerous and limited manual detection of underwater structures,and it can be used for underwater structure surface disease detection for the actual project.