عنوان مقاله [English]
Hydraulic engineering studies need to know the sizes of surface sediments exposed to flow. The conventional ways, manual measurement methods such as performing gradation tests (sieve analysis) in a laboratory, face many problems in determining the sizes of sediments. An algorithm has been developed to facilitate problems. By providing a number of images of the bed materials during field surveys and providing land use information, this algorithm can help accelerating processing and increasing the accuracy of modeling. To verify the accuracy and precision of this algorithm, versus the conventional soil gradation method, four different digital cameras (with 6, 10, 12.1, 14 megapixels resolution) and two square-frames (0.5 and 1 meter in size) used for imaging and sampling bed materials, in dry and under low water flow conditions, in a distance of 5 km in Sirch River, north east of Kerman. The results showed that cameras with high resolution had the least error in estimating the average diameters of materials in dry beds. However, due to the uniformity of the bed surface particles, the performance of cameras with 6 and 14 megapixels was also acceptable, so that the maximum difference in the average diameter estimation was less than 12%. The cameras with 12.1 and 10-megapixels had lowest error in estimation the maximum particle's diameters, respectively. Furthermore, calculations carried out in the wet bed surface indicated that the results of images processing obtained from the camera with a low resolution, 6 megapixels, differ greatly from what had been obtained by other cameras, especially in distribution of the gradation curve and statistical characteristics of the particles. A survey on the statistical characteristics of the bed surface materials distribution showed that the average diameter, maximum diameter and deviation of the particles obtained from the camera with a resolution of 10 megapixels were greater than those obtained by other cameras. In a general conclusion, it can be stated that there is a direct relationship between these parameters: the image dimensions (frame), the resolution of the camera and its distance from the stream bed, in a way that by reducing the resolution of the camera in a fixed frame, the camera's distance from the stream bed should be increased, and vice versa. Based on the results in the streams with fairly uniform grain distribution, any camera can be used. But in streams that have a coarse-grained or fine-grained bed, it's better to use low and high resolution cameras, respectively.