Hyperspectral remote sensing images (HSIs) have been applied in urban planning, environmental monitoring, and other fields. However, they are susceptible to noise interference, such as Gaussian noise, stripe, and mixed noises, from various factors in the imaging process, which greatly limits their applications. Although previous efforts to improve HSI quality have achieved remarkable results, there are still many challenges to be solved. To avoid the poor generalization ability and improve the stripe removal performance of the network in real scenarios, in this article, we proposed a novel deep learning model (Translution-SNet) for HSI stripe noise removal based on a semisupervised training strategy that applies a convolution and transformer for feature extraction. Moreover, we used an unbiased estimation method to calculate the loss function of the unsupervised part from noisy data without a clean image. The semisupervised method improved the ability of Translution-SNet to deal with various complex stripe noises during stripe removal and strengthened its robustness and generalization ability. Our experimental results showed that Translution-SNet could robustly handle stripe noise of images with different loads and achieve satisfactory results, proving its feasibility and effectiveness. In addition, Translution-SNet showed good generalization ability.

The traditional profile model of the sediment diffusion coefficient (), which is parabolic along the water depth, cannot precisely predict the vertical distribution of in vegetated flow. Therefore, in this study, a series of flume experiments were conducted to clarify the influence of submerged vegetation on the diffusion characteristics of suspended load particles. First, the submerged canopy increases the efficiency of momentum exchange but restrains the vertical diffusion of suspended sediment, indicating that the presence of submerged plants promotes the siltation of solid particles. Second, the depth‐averaged sediment diffusion coefficient () monotonically decreases with increasing relative water depth, while the depth‐averaged momentum exchange coefficient () presents two opposite trends, likely owing to the dominant turbulence event changing from ejections to sweeps and the different variation tendencies of the velocity gradient and Reynolds shear stress with increasing vegetation density. Compared with the ratio () of to the momentum exchange coefficient in nonvegetated flow, in submerged canopy flow is significantly less than 1, which means that the solid particles diffuse less readily than liquid particles because of the stem‐scale vortices generated by the vegetation. In addition, vertical profile models of and suspended sediment concentration were proposed and validated by experimental data. The models exhibit a high accuracy, correlation and applicability and can provide critical information to promote research on riverbed deformation and nutrient dynamics in vegetated rivers, wetlands and estuaries.

Vegetation in natural river interacts with river flow and sediment transport. This paper proposes a two-layer theoretical model based on diffusion theory for predicting the vertical distribution of suspended sediment concentration in a flow with submerged vegetation. The suspended sediment concentration distribution formula is derived based on the sediment and momentum diffusion coefficients through the inverse of turbulent Schmidt number ( $S{c}_t$ ) or the parameter $\eta$ which is defined by the ratio of sediment diffusion coefficient to momentum diffusion coefficient. The predicted profile of suspended sediment concentration moderately agrees with the experimental data. Sensitivity analyses are performed to elucidate how the vertical distribution profile responds to different canopy densities, hydraulic conditions and turbulent Schmidt number. Dense vegetation renders the vertical distribution profile uneven and captures sediment particles into the vegetation layer. For a given canopy density, the vertical distribution profile is affected by the Rouse number, which determines the uniformity of the sediment on the vertical line. A high Rouse number corresponds to an uneven vertical distribution profile.