초분광 CCTV 기반 부유사 계측 적용성 평가: 감천 유사 추적자 실험 사례

초록

Suspended sediment plays a key role in river morphodynamics and water quality, yet conventional in-situ measurements limit continuous, high-resolution monitoring. This study evaluates the applicability of CCTV-Hyperspectral Imaging for Suspended Sediment Transport (CCTV-HISST) for measuring suspended sediment concentration (SSC) in a natural river. We compare tracer experiments using fine sediment in Gam Creek under varying solar irradiance conditions with results from an outdoor full-scale flume experiment at the River Experimental Center (REC). To reduce uncertainties caused by natural light variability, the CCTV-HISST system applies a real-time radiometric correction scheme based on simultaneous observations of the water surface and a Spectralon reference panel. We estimates SSC using the Light Classification and Adaptive Regression (LiCAR) framework coupled with a Random Forest model and achieves high predictive accuracy (R2 = 0.97, MAPE = 13.3%) with minimal bias, comparable to the REC experiments. Gam Creek and REC data show similar overall spectral sensitivity patterns. However, differences in water depth, sediment mixing, and optical conditions introduce systematic shifts in spectral-SSC relationships, with increased importance of near-infrared wavelengths (~850 nm). The analysis further extends surface SSC map vertically using the Rouse-alpha model to reconstruct continuous three-dimensional SSC fields. These results demonstrated that CCTV-HISST provides robust SSC monitoring in natural rivers under complex hydraulic and optical conditions and highlight its potential for practical river monitoring and hydraulic-environmental management applications. © 2026 Korea Water Resources Association.

키워드