Parametric image-based concrete defect assessment method

Structural health monitoring aims to ensure the integrity of infrastructure. Assessing structural integrity through image classification techniques based on human perception is often challenging. Incorporating a morphological image-based crack detection algorithm can mitigate these limitations. This study proposes a parametric approach for concrete crack analysis using digital image processing techniques. The objective is to detect damage to the concrete surface of infrastructures, potentially exacerbating structural deterioration. The research justifies the utilization of digital parametric image processing techniques, including a weighted median filter, grayscale, Otsu filter, and other methods. These techniques were applied to evaluate different concrete images for crack detection. The findings were validated through Image Quality Assessment (IQA), determining the condition of the studied images and statistical properties using metrics such as Mean Squared Error (MSE), Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index (SSIM), Blind/Referenceless Image Spatial Quality Evaluator (BRISQUE), Natural Image Quality Evaluator (NIQE), and Perceptual Image Quality Evaluator (PIQE). Additionally, uncontrollable factors such as variations in lighting effects under different conditions, testing setups, and different textures of the concrete were considered in this study. The comprehensive experimental results demonstrated that the Otsu filter outperforms other filtering techniques. Overall, the study achieved remarkable accuracy, approximately 95%, in detecting cracks on building infrastructure. The proposed method holds potential for integration with advanced techniques in practical applications for the maintenance and safety of infrastructure. This research not only contributes to technological advancements in concrete defect assessment but also carries significant implications for the future of automated and reliable structural health monitoring.