Introducing and Evaluating the Effectiveness of Various Image Processing Algorithms in Determining Hydraulic Raughness Using Gradation Curve in Gravel Bed Rivers

The importance and role of bed roughness in conducting hydraulic flow studies and sediments, especially in river simulation models, are not covered by any of the experts in this area. For this reason, researchers used simple methods such as direct measurement using ruller or calipers or mechanical sieving for many years to determine the size of roughness of bed particles. In recent years, with the advancement of technology and the promotion of digital photography cameras, various algorithms and software for automated analysis of bed surface data are provided using digital images taken from the river bed. But so far, no comparative study has been conducted regarding the comparison of the results of different methods and their validation and only researchers have been evaluating and commenting on the shape of the gradation curve and the particle diameters. In this study, imaging of one kilometer from the bed of the Kordan river in the west of the Alborz province, and then the images obtained in nature, as well as images taken from the particles collected from the bed in the laboratory with three different image processing methods and in two backgrounds with different colors have been analyzed and compared with the results of the sieve analysis method. Based on the results, it can be said that the presentation of an acceptable gradation curve does not mean the accuracy of image methods, and these methods are sometimes found to have a notable error in the counting of particles, which, by shrinking the image field or increasing the number of fine particles in the image or the presence of particles with longitudinal elongation, this error will be tangible and indescribable. Factors such as how adjacent particles are placed, the color and appearance of aggregates, the distribution of light in the image, and the algorithm used to process the image on the grain size curve, are effective. Therefore, using these methods in estimating hydraulic roughness
should be done with sufficient accuracy.