It is very hard to access cloud-free remote sensing data, especially for the ocean color images. A cloud removal approach from ocean color satellite images based on numerical modeling is introduced. The approach removes cloud-contaminated portions and then reconstructs the missing data utilizing model simulated values. The basic idea is to create the relationship between cloud-free patches and cloud-contaminated patches under the assumption that both of them are influenced by the same marine hydrodynamic conditions. Firstly, we find cloud-free GOCI (the Geostationary Ocean Color Imager) retrieved suspended sediment concentrations (SSC) in the East China Sea before and after the time of cloudy images, which are set as initial field and validation data for numerical model, respectively. Secondly, a sediment transport model based on COHERENS, a coupled hydrodynamic-ecological ocean model for regional and shelf seas, is configured. The comparison between simulated results and validation images show that the sediment transport model can be used to simulate actual sediment distribution and transport in the East China Sea. Then, the simulated SSCs corresponding to the cloudy portions are used to remove the cloud and replace the missing values. Finally, the accuracy assessments of the results are carried out by visual and statistical analysis. The experimental results demonstrate that the proposed method can effectively remove cloud from GOCI images and reconstruct the missing data, which is a new way to enhance the effectiveness and availability of ocean color data, and is of great practical significance.
The distribution and transport of suspended sediment in the coastal areas has attracted more and more attention. Monitoring and modeling of the distribution and transport of suspended sediment is significant. The mutual promotion of remote sensing and numerical simulation plays an important role on the coastal water quality study. In this study, a method of coupling derived suspended sediment concentration (SSC) images with numerical model, based on GOCI and COHERENS (COupled Hydrodynamical-Ecological model for REgioNal and Shelf seas) model, is proposed to monitor the suspended sediment dynamics in the East China Sea. The retrieved SSC were extracted from GOCI images, to set as initial condition and employed to calibrate the parameters and validation for the hydrodynamic modeling and sediment transport modeling, respectively. The model is forced by considering tidal surface elevation at open sea boundary, river discharges, surface stress as a function of wind speed, air temperature, relative humidity and cloud coverage, bottom roughness and heat flux through sea surface. The model results are in accord with the in situ measurements. The results show that: (1) Numerical model which initialized with satellite-derived SSC data can quickly response to the changes of sediment concentration in real sense. (2) Remote sensing is helpful to calibrate and validate the model for simulating the suspended sediment concentration distribution. (3)The proposed approach can obtain reasonable simulated results in the East China Sea. (4) It is of great significance to combine remote sensing and numerical simulation together to study the water quality in the coastal areas.
Lake ecological environment is changing, driving by natural and human factors, and in turn influence people's living and
producing. Therefore, dynamic monitoring of lake based on remote sensing technologies will play an important role to the
disaster prevention and reduction work of lakes. In this paper, we expounded a series of work to realized monitor Poyang
Lake dynamically by using HJ-CCD images. First, we did pretreatment to all HJ-CCD images, which mainly contain
geometric correction, atmospheric correlation, image clipping, etc. Then, based on different features between water and
non-water in different index layers, we extracted the covered area by water in different times from the corresponding
HJ-CCD images, and we also extracted the true area through visual interpretation method. After that, by combining the
water boundaries and DEM, we also estimated water level and water capacity in different times. Results of our work
showed that the mean absolute error of water area extracted through remote technologies is 5.57%. The relationship of
remote sensing areas and visual interpretation areas could be described as Strue = 0.8757*Sinterp + 110.24, with R2 = 0.9807.
Besides, there was obvious relationship between water area and water capacity of Poyang Lake too, and the relations can
be described with linear function. Based on such results, we can realize the dynamic estimation of Poyang Lake’s area and
capacity from daily gotten HJ-CCD image which covers the District of Poyang Lake. In other words, the results of this
paper can provide decision basis for Poyang Lake’s real-time, dynamic, economic monitoring.