A Study of Image-Based Algorithms of River Sediment Measurement

Letting an underwater camera sunk into a pre-set vertical location by using the existing hydrology cableway in river, When the computer in the bank workstation issued camera control commands, the underwater camera begins taking photo to water samples, and the image transmitted to the image acquisition card, and transmitted the image data which be collected by ARM to the bank workstations. Various pieces of image be denoising, segmentation, calculation and processing by riparian workstation computer, and thus arrive at the river sediment content. This proposed method image-based content of river sediment detection techniques applied to the lower sediment concentration (30 kg/m³) of river sediment measurements and used in hydrological stations of Wujiang City, Ji´an practical, the application results show that the method than the traditional drying method in time, a decrease of 96% accuracy compared with the drying method error of 1%-5%, in line with norms of hydrologic and sediment measurements. Although the method for the large sediment (100 kg/m³) river sediment concentration measurement still falls short of expectations, but it can still test method used for the design and optimization to provide a frame of reference.