Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Fast Measuring Pavement Crack Width by Cascading Principal Component Analysis

Zhicheng Wang, Junbiao Pang

TL;DR

This work addresses the challenge of quantifying pavement crack width from images, which is difficult due to irregular crack boundaries and the need for rapid, point-wise measurements. It introduces Cascade PCA (CPCA), a three-stage, unsupervised framework that first segments cracks, then uses PCA for low-complexity cracks and RPCA for high-complexity cracks to estimate the Main Propagation Axis (MPA) and measure width by rotation to the vertical axis. The key contribution is the rejection chain that balances speed and accuracy by applying RPCA selectively, supported by strong empirical results on CFD, Cracktree, and Crack500 showing superior MAE/MSE performance and favorable running times. This approach enables robust, automated crack width quantification suitable for scalable pavement condition assessment with minimal parameter tuning.

Abstract

Accurate quantification of pavement crack width plays a pivotal role in assessing structural integrity and guiding maintenance interventions. However, achieving precise crack width measurements presents significant challenges due to: (1) the complex, non-uniform morphology of crack boundaries, which limits the efficacy of conventional approaches, and (2) the demand for rapid measurement capabilities from arbitrary pixel locations to facilitate comprehensive pavement condition evaluation. To overcome these limitations, this study introduces a cascaded framework integrating Principal Component Analysis (PCA) and Robust PCA (RPCA) for efficient crack width extraction from digital images. The proposed methodology comprises three sequential stages: (1) initial crack segmentation using established detection algorithms to generate a binary representation, (2) determination of the primary orientation axis for quasi-parallel cracks through PCA, and (3) extraction of the Main Propagation Axis (MPA) for irregular crack geometries using RPCA. Comprehensive evaluations were conducted across three publicly available datasets, demonstrating that the proposed approach achieves superior performance in both computational efficiency and measurement accuracy compared to existing state-of-the-art techniques.

Fast Measuring Pavement Crack Width by Cascading Principal Component Analysis

TL;DR

This work addresses the challenge of quantifying pavement crack width from images, which is difficult due to irregular crack boundaries and the need for rapid, point-wise measurements. It introduces Cascade PCA (CPCA), a three-stage, unsupervised framework that first segments cracks, then uses PCA for low-complexity cracks and RPCA for high-complexity cracks to estimate the Main Propagation Axis (MPA) and measure width by rotation to the vertical axis. The key contribution is the rejection chain that balances speed and accuracy by applying RPCA selectively, supported by strong empirical results on CFD, Cracktree, and Crack500 showing superior MAE/MSE performance and favorable running times. This approach enables robust, automated crack width quantification suitable for scalable pavement condition assessment with minimal parameter tuning.

Abstract

Accurate quantification of pavement crack width plays a pivotal role in assessing structural integrity and guiding maintenance interventions. However, achieving precise crack width measurements presents significant challenges due to: (1) the complex, non-uniform morphology of crack boundaries, which limits the efficacy of conventional approaches, and (2) the demand for rapid measurement capabilities from arbitrary pixel locations to facilitate comprehensive pavement condition evaluation. To overcome these limitations, this study introduces a cascaded framework integrating Principal Component Analysis (PCA) and Robust PCA (RPCA) for efficient crack width extraction from digital images. The proposed methodology comprises three sequential stages: (1) initial crack segmentation using established detection algorithms to generate a binary representation, (2) determination of the primary orientation axis for quasi-parallel cracks through PCA, and (3) extraction of the Main Propagation Axis (MPA) for irregular crack geometries using RPCA. Comprehensive evaluations were conducted across three publicly available datasets, demonstrating that the proposed approach achieves superior performance in both computational efficiency and measurement accuracy compared to existing state-of-the-art techniques.

Paper Structure

This paper contains 17 sections, 13 equations, 8 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Pixel-Wise Crack Detection
  4. Crack Width Measurement
  5. Cascade PCA
  6. Problem Definition
  7. PCA for the Low Complexity Crack
  8. RPCA for High-Complexity Cracks
  9. Group PCA and RPCA into A Rejection Chain
  10. Experiments
  11. Database
  12. Evaluation
  13. Baselines and The State-of-Art Approaches
  14. Comparisons with the SOTA methods
  15. Running Time
  16. ...and 2 more sections

Figures (8)

  • Figure 1: A crack patch is rotated to parallel the Main Propagation Axis (MPA) of the crack. The green point is the width measurement point we selected, the red line $a$ is MPA of the crack, and the blue line $b$ is perpendicular to the red line.
  • Figure 2: (a) is a crack and (b) is the crack skeleton corresponding to (a).
  • Figure 3: The yellow and red pixels are the check points we selected; the green lines are the edge lines which correspond to the check points, and the red line is the MPA (best viewed in color).
  • Figure 4: PCA and RPCA are organized into rejection chain
  • Figure 5: The green line fitted by RANSAC almost parallels to the MPA of a crack (in red). The red pixel is the check pixel (best viewed in color).
  • ...and 3 more figures

Theorems & Definitions (2)

  • Definition 1: Check Point
  • Definition 2: MPA