Adaptive water waste processing strategy at floating barriers using computer vision, route finding, and Monte Carlo simulation

Abstract

Floating barriers installed along the riverbanks block floating debris (water trash) from contaminating marine environments. Regular removal of accumulated trash before reaching weight capacity is crucial to maintain structural integrity and prevent trash overflow. However, research on the costs of collecting floating debris in water infrastructure has been insufficient. To fill this knowledge gap, this study investigates the costs of processing water trash at multiple floating barriers and presents a novel water trash processing framework comprised of trash detection, valuation, and collection planning. The proposed framework (1) detects the types and mass of collected trashes using computer vision, (2) evaluates the process cost of the water trashes, and (3) derives an optimal garbage collection path planning. Monte Carlo Simulation is employed to simulate water trash collection and processing scenarios for estimating the total associated costs. Experimental results showed that the proposed framework achieved a 10% to 30% cost reduction compared to conventional time-based collection methods. The proposed water trash processing framework and the findings will contribute to our understanding on the costs of processing water trash at floating barriers to prevent ocean pollution, thereby facilitating the implementation of such infrastructure and planning the budgets required for their operation and maintenance.