Image-Based Automated Width Measurement of Surface Cracking.

Area of Science:

• Civil Engineering
• Materials Science
• Computer Vision

Background:

• Traditional crack detection relies on manual visual inspection and crack-width comparator gauges (CWCG), which are time-consuming, subjective, and prone to measurement errors.
• Existing automated crack detection algorithms often focus on segmentation using Deep Learning but neglect the critical aspect of accurate crack width evaluation.

Purpose of the Study:

• To propose and validate a novel automated method for precise surface crack width measurement using digital image processing.
• To develop a technique that addresses the limitations of traditional methods and current automated approaches by focusing on width evaluation and orientation.

Main Methods:

• The proposed method involves three stages: anisotropic smoothing for noise reduction, image segmentation to isolate cracks, and k-means adjustment for stabilizing central points.
• This approach enables the characterization of both crack width and curvature-related orientation.

Main Results:

• The method was validated on fiber-reinforced earthen construction materials, demonstrating robustness, efficiency, and high accuracy in estimating crack width from digital images.
• It successfully distinguishes real cracks from false ones and can detect cracks as narrow as 0.15 mm, irrespective of lighting conditions.

Conclusions:

• The developed digital image processing technique offers a significant advancement in automated crack width measurement for civil infrastructure.
• This novel approach provides a reliable and efficient tool for assessing structural safety and integrity, overcoming the drawbacks of manual inspection and existing automated methods.