The Research On X-ray Film Of Oil Pipeline CCD Image Processing Method

The X-ray inspection technology with film image is an important method ofindustrial non-destructive testing. It plays an irreplaceable role in detecting the defect ofthe oil pipelines. The artificial rating, although is mostly used, is low efficiency, pooraccuracy and has large health hazard for the person use it. It has significant value thatstudying the method of applying modern digital image processing technology on X-rayimages, improving the accuracy and efficiency of defect detection. In this dissertation,research is focused on the digitization and image processing methods of digitized for theoil pipeline’s X-ray image. The main content of the dissertation is summarized asfollows:(1). According to the disadvantage of serious noise and low contrast in the X-rayfilm image, the source of noise and noise model are analyzed. Through the theoreticalanalysis and experimental contrast of mean filter, median filter and Gaussian filter, themethod of X-ray film image de-noising using33window median filter to removepulse noise and using55window Gaussian template filter to remove Gaussian noiseare proposed.(2). The method for image enhancement based on the histogram and gray-scaletransformation are studied. By the theoretical analysis and experimental comparativestudy, the conclusion, which histogram equalization is not suitable for enhancementpending X-ray film image nevertheless the logarithmic transformation or powertransformation is suitable for enhancing the image, is obtained. And a new method ofX-ray film image enhancement based sine function is proposed.(3). The method of the edge detection based on gradient template or Cannyguidelines is studied，as well as the threshold-based image segmentation method.Through the theoretical analysis and experimental contrast, the conclusion that cannyguidelines detecting edge need to study the adaptive algorithm and defect segmentationbased on Otsu algorithm need to reduce the interference information is got. And themethod of adaptive Canny defect extraction based on Otsu is proposed.(4). By theoretical analysis and experimental comparative study of the basic morphological operations and structural elements, the method of X-ray film imagedefects’ edge extraction based on an anti-noise dilation and erosion morphologicalgradient is proposed(5). A hardware system of X-ray film image defects extraction is designed. Bysummarizing theoretical analysis and experimental comparative study of imagede-noising, enhancement, feature extraction and segmentation, two schemes of edgeextraction, using image enhancement based on the sine function and adaptive Cannymethod based on Otsu or using logarithmic function enhancement and morphologicalgradient of anti-noise dilation and erosion type, are proposed. The experimental resultsshow the effectiveness of the defect edge extraction.