A refined automated grain sizing method for estimating river-bed grain size distribution of digital images

► Image processing techniques facilitate automated grain size measurement. ► We propose a refined automated grain sizing method (R-AGS) for estimating grain size. ► 130 Digital images are used to assess the R-AGS performance. ► The R-AGS outperforms two pivotal methods in estimating grain size distributions. ► This study moves one step toward the practical automated grain-size measurements. Natural bed topography and habitat is affected by the composition of gravels in various shapes and sizes. Traditional measurement methods for grain size distribution are time-consuming and labor-intensive. Recent advances in image processing techniques facilitate automated grain size measurement through digital images. This study introduces a refined automated grain sizing method (R-AGS) incorporating a neural fuzzy network for automatically estimating the grain size distribution, specifically for digital images composed of grains ranging from 16mm to 512mm. A total of 130 digital images captured from the Lanyang river-bed in northeast Taiwan are used to assess the R-AGS performance. We demonstrate the neural fuzzy network can adequately identify the binary threshold, which is a crucial parameter of the AGS procedure, and the proposed R-AGS can be intelligibly used for automated accurate estimation of grain size distribution with much less labor-intensiveness for each digital image. Moreover, it is easy to re-construct the network by updating rule nodes for image samples significantly different from this study; consequently its applicability and practicability could be expanded.