Theory of belt-oriented RADON transform is developed, and applied to extracting features from high-resolution remote sense images. By several typical experiments it is proved that belt-oriented Radon transform is powerful in extracting belt features, but Radon transform (line-oriented Radon transform called by author) is unavailable.
Remote sensing is an indispensable means for coastal band monitoring. Using satellite remote sensing data to monitor coastline variation and to analysis the eroding, depositing features and evolution process will be of great significance for the river mouth regulation, river course planning, coastal protective project program and trend prediction of coastal evolution. So, it is necessary to establish a coastline dynamic monitoring system. The system is mainly based on remote sensing and spatial information analysis techniques. In this paper, the system framework, design methodology and system functions are described in detail. The key techniques and methods involved in the system construction are particularly discussed, and they include the data preprocessing techniques, such as cloud identification and geometry fine correction, multi-scale coast edge extracting algorithm based on MRF model and coastline tide correction model with measured data or numerical simulation result as input, and coastline dynamic analysis method based on time series analysis and spatial topological analysis. Finally, an example to apply the system to the Yellow River mouth delta is given and the process flow diagram and procedures are described. The comparison of monitoring results with manually interpreted results has verified the favorable effect of the system.