THREE LAYER MEDIAN FILTER METHOD FOR IDENTIFYING CONCRETE STRENGTH LEVELS BASED ON CONCRETE IMAGES

This study introduces a novel approach utilizing digital image processing techniques to analyze concrete surface images for categorizing concrete strength levels based on two-dimensional RGB digital photographs. The research addresses the limitations of traditional median filters, such as insufficient noise reduction and edge preservation, by proposing a three-layer median filter for enhanced image preprocessing. The methodology involves three main phases. First, RGB images are converted to the Lab color space, followed by segmentation using the K-Means clustering method and noise reduction through the proposed three-layer median filter. This approach improves noise suppression by 15% compared to traditional median filters, as verified through quantitative analysis. Second, shape and texture features are extracted from the processed images to capture distinctive characteristics of the concrete surface. Finally, the images are classified into strength levels ranging from K100 to K300 using these features. The proposed method achieved a 90% accuracy rate, correctly identifying 46 true positives (TP) and 44 true negatives (TN), with minimal errors from 6 false negatives (FN) and 4 false positives (FP). This represents a significant improvement over conventional methods. The findings validate the robustness and reliability of the proposed method in accurately classifying concrete strength levels. By addressing key challenges in traditional approaches and integrating advanced image processing and clustering techniques, this research provides a non-destructive and efficient alternative for evaluating concrete strength. The study establishes a foundation for future advancements in automated material characterization and quality control in construction and engineering domains.