The performance of gravel bedload predictor equations is particularly poor.A key reason for this situation is the current poor understanding of the physical mechanisms that are involved in the entrainment of grains from water-worked sediment deposits.This has been caused,in part,by a lack of detailed data on the behaviour of moving sediment grains.In this study,observations of the entrainment, transport and deposition of many particles from a gravel bed has been collected and studied using an image-based system.This system was used to measure the motion of hundreds of gravel particles moving as bedload.This paper describes the development of the image-based system and the data processing methods used to provide the automated measurement of particle motion.This system was adapted from commercially available PIV equipment and was configured to obtain simultaneously data for the analysis of the moving grains and the near-bed flow field.The techniques developed are relatively sensitive to sediment colour and shape by using related image processing algorithms and so could be applied to natural gravel sediments.The results(image data)obtained from the techniques developed are amended in a program by using two functions(commands)of image processing called 'Label2RGB',and 'RegionProps' which distinguish the colour and shape of moving sediment grains respectively.The data processing provided information that identified individual grain entrainments,depositions.We accurately traced the motion of some,but not all,particles.Results show that the system was able to estimate grain entrainment rates and indicate locations of preferential entrainment and deposition from data on hundreds of individual moving grains.
展开▼
