Structural integrity is a paramount concern in engineering, and crack detection plays a crucial role. This article compares image processing techniques for crack detection in structures. We explore image transformations such as inversion, grayscale conversion, and logarithmic and power transformations to highlight crack features. The effectiveness of renowned edge detection operators like Roberts, Prewitt, Sobel, and Laplace, is investigated, and their performance evaluated based on edge detection quality and noise resistance. Thresholding methods, particularly the iterative algorithm and OTSU method, are discussed. These methods convert grayscale images into binary representations, simplifying crack identification by isolating crack regions from the background. Our techniques are validated on real-world images of structurally compromised components, and evaluated using metrics such as accuracy, precision, and computational efficiency. We also include a practical code snippet to demonstrate the image processing techniques, from loading the original image to generating a boundary image. This research provides a comprehensive comparative analysis of image processing techniques for structural crack detection, aiding engineers and researchers in method selection based on their specific requirements. Additionally, our techniques open avenues for further investigation into automating these techniques, ensuring continuous structural health monitoring. This can significantly enhance crack detection in structural components, facilitating timely maintenance and ensuring structural safety and longevity.
Anahtar Kelimeler: Structural Crack Detection, Image Processing Techniques, Edge Detection Operators, Thresholding Methods, Structural Integrity and Maintenance
|
