Characteristic analysis of floc size distribution and image texture evolution in chemical coagulation process

The dynamic evolution characteristics of flocs during chemical coagulation play a critical role in water treatment process optimization, yet real-time monitoring of reaction conditions and floc morphology remains challenging. This study innovatively integrates machine vision with image texture analysis to systematically investigate the regulatory mechanisms of coagulant dosage and raw water turbidity on floc size distribution and textural features. A non-invasive high-speed imaging system was employed to capture the entire flocculation dynamics, coupled with Python-OpenCV algorithms for quantitative characterization of floc parameters. Results demonstrate that under constant turbidity: (1) Insufficient coagulant dosage leads to inadequate colloidal destabilization, manifested by reduced floc quantity and significant decreases in image gray mean, entropy, and correlation values; (2) Optimal dosage produces concentrated floc size distribution with low coefficient of variation (CV); (3) Overdosing induces floc erosion and fragmentation, forming bi-/tri-modal distributions with substantially increased CV. Furthermore, high-turbidity raw water exacerbates floc fragmentation, resulting in dispersed size distribution (elevated CV) and enhanced light scattering (improved texture contrast). This work first elucidates the synergistic evolution between image textural features and floc size distribution, confirming their potential as sensitive indicators for real-time coagulant optimization, thereby providing theoretical foundations for intelligent control of coagulation processes.