STUDY ON USING IMAGE ANALYSIS TO DETERMINE THE RELATIONSHIP BETWEEN BOOM BEHAVIOR AND ENVIRONMENT IN A RESERVOIR

Abstract

 In recent years, the number of combined disasters caused by floods and driftwood has been increasing, and countermeasures against driftwood disasters are required. Dam reservoirs have the role of trapping driftwood with a boom, which contributes to reducing flood damage downstream. Although booms play such an important role, the age of booms is unclear, and the timing of their replacement can be determined by appearance. In view of the increasing external force of floods and the amount of inflowing wood, it is important to clarify the deterioration mechanism of booms and to maintain the booms to prevent it from breaking during floods. In this study, we proposed a boom detection method using image processing that is robust against noise and obstacles, in order to observe the behavior of booms during normal and flood conditions from time-lapse images of the booms taken at the Amagase Dam, and to clarify the relationship with hydraulic and meteorological observations. In addition, the correlation between the behavior of the detected net field and the observed wind direction, wind speed and flow rate is investigated.