Application of an image processing-based algorithm for river-side granular sediment gradation distribution analysis

Determining grain-size and grading distribution of river-side sediments is very important for issues related to lateral embankment drift, river-side nourishment, management plans, and riverbank stability. In this regard, experimental procedures such as sieve analysis are used in regular assessments which require special laboratory equipment that are quite time consuming to perform. The presented study provides a machine vision and image processing-based approach for determining coarse grained sediment size and distribution that is relatively quick and effective. In this regard, an image image processing-based method was used to determine the particle size of sediments as justified by screening tests which were conducted on samples taken from the riverside granular sediments. As a methodology, different grain identification stages were applied to extract sediment features such as pre-processing, edge detection, granular size classification and post-processing. According to the results of the grain identification stages, the applied technique identified about 35% sand, 55% gravel and 7% cobble which is approximately near to the screen test results which were determined as 30% sand, 52% gravel, and 5% cobble. These results obtained from computer-based analyses and experiments indicated that the utilised processing technique provided satisfactory results for gradation distribution analysis regarding riverside granular sediments.