Quantifying the Surface Texture of Coarse Aggregate Through Enhanced Image Processing

In civil engineering and material science, surface texture analysis of construction materials is crucial for predicting mechanical properties and performance characteristics of concrete and asphalt mixtures. While existing methods provide valuable insights, they face limitations including computational complexity, noise sensitivity, expensive setup, and manual intervention requirements. Current study presents an innovative image processing technique for quantifying aggregate surface texture through a methodology combining multi-angle pixel intensity analysis and Gray-Level Co-occurrence Matrix (GLCM) features. The approach introduces three key innovations: a robust preprocessing pipeline including edge detection, color conversion, noise reduction, and texture quantization; multi-angle pixel intensity analysis at 0°, 45°, 90°, and 135° to capture comprehensive texture information; and an automated maximal value selection algorithm that optimizes texture measurement while minimizing noise. Experimental validation demonstrated excellent reliability with texture measurements varying less than 1.5% across different image capturing devices and images taken at different heights. The program distinguished various surface textures, from smooth surfaces (texture value ~0) to rough aggregate surfaces (texture values >1.5), offering a more accessible, automated, and accurate approach to texture quantification which can be applicable in material characterization, and quality control applications.