Dissertation > Excellent graduate degree dissertation topics show

The Research on the Extraction of the Weld Flaw Images

Author: YangChuan
Tutor: GaoZuo
School: Wuhan University of Technology
Course: Marine electric propulsion and automation technology
Keywords: Gaussian filtering Fuzzy Enhancement Binarization Mark defects Edge tracking
CLC: TP391.41
Type: Master's thesis
Year: 2010
Downloads: 214
Quote: 4
Read: Download Dissertation

Abstract


The ship welding is an important factor of the ship in ship construction quality is good or bad, the quality of welding a direct impact on the durability and safety of the ship is an important prerequisite to ensure the quality of ships, safety of navigation and safe operation. Therefore, the the ship welding defects detection is a particularly critical step in the process of ship construction. Vessel weld inspection in the most commonly used X-ray radiography, this detection method to produce large amounts of weld film, however, most of these films are based on artificial assessment approach, due to the overall quality of the piece of a staff technical experience impact of artificial evaluation method of the heavy workload, low efficiency, accuracy is not high. Computer-aided evaluation chip technology, greatly reducing the workload of the evaluation sheet, improve work efficiency, and a piece of a process more scientific and standardized. However, due to the imaging characteristics of the weld film image of the weld image there is too much noise, the defect edge blur and low contrast, weld defects correct extraction, feature calculation and subsequent classification and recognition has led to great difficulties. The film image processing problems for the computer-aided evaluation, the paper in-depth study and analysis of the seam image processing process, including the following: First, the process for the ship welds film images to obtain particularity, focusing on analysis of the digital features of the image, through comparative studies appropriate image processing method. Starting from the X-ray film image forming process, the imaging characteristics of the film and the digitization process several different to summarize obtained welds image characteristics by the gradation characteristics of the transformed digital image analysis, i.e. the complex and mainly Gaussian noise The noise, the edge of the weld area and defect edge blur, and weld area and defect contrast is not high; then use these characteristics different denoising and enhancement algorithms for comparative analysis, get the best approach to Gaussian smoothing and fuzzy enhancement; Secondly, the defect marking and boundary tracking algorithm defects in the weld on the basis of the image preprocessing, the distinction mark and the edge of the statistical information. Eight-connected labeling algorithm and boundary tracing algorithm extracted from the individual defect pixels defects defective tags and boundary storage, in order to facilitate the weld defect characteristics; Finally, the defects of geometric characteristics of quantitative calculation. By Mark defects and edge extraction based on the marked pixel calculation and calibration conversion to give each defect in the weld area, perimeter and minor axes of three geometric characteristic values, obtain a more satisfactory results for the correct defect classification and defect rating ready for work.

Related Dissertations

  1. Research on License Plate Location Algorithm of License Plate Recognition System,TP391.41
  2. Art portrait paper-cut generation and psychological cognitive appraisal,TP391.41
  3. Facial feature points to generate automatic annotation and expression,TP391.41
  4. Based on digital image processing meter image recognition technology research and implementation,TP391.41
  5. Low-quality document image binarization research,TP391.41
  6. Research on Plate Positioning Algorithm Based on Image Processing,TP391.41
  7. Research on Real-Time License Plates Segmentation and Recognition Techniques,TP391.41
  8. Fingerprint recognition algorithm research related,TP391.41
  9. Research on Face Recognition Based on Improved Maximum Margin Criterion,TP391.41
  10. Research on Palm Vein Image Acquisition and Thining Technology,TP391.41
  11. Correction Algorithm for Fiber Images with Non-Uniform Illumination,TP391.41
  12. Binarization and Its Application in Preprocessing of Cheque Recognition System,TP391.41
  13. Study Runway Recognition Based on Airport Remote Sensing Image,TP751
  14. Research and Design of Korean License Plate Detection System,TP391.41
  15. Video text extraction technology research and application,TP391.41
  16. A Research for Three Dimensional Object Reconstruction,TP391.41
  17. Research on the Main Technology of a Vehicle License Plate System,TP391.41
  18. Design and realization of fingerprint recognition in image processing method,TP391.41
  19. Based on Object-oriented Programming Graph Arrow Quantification Research,TP391.41
  20. Tire Defects Inspection Based on Digital Image Processing,TP274
  21. The Study of Boundary Detecting Algorithm Based on Joint Entropy,TP311.13

CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Pattern Recognition and devices > Image recognition device
© 2012 www.DissertationTopic.Net  Mobile