The Research Of Object-Based Remote Sensing Change Detection For Coastal Surface

Coastal areas are located in the transitional zone between lands and seas, where abundant resources have provided a firm material foundation for human society. With the rapid economic development and excess population growth, coastal Land-Use and Land-Cover Changes are accelerating as well as significant changes in coastal environment. It is difficult to meet monitoring coastal changes by regular coastal survey means as complex coastal environment,inadequate data and difficult data processing. Satellite remote sensing technology for exploring the Earth is one of the best means for monitoring coastal changes and management. With the launch of high resolution sensors, the range of applications of high resolution remote sensing data is expanding, the form of applications of high resolution remote sensing data is varied and the level of applications of high resolution remote sensing data is also improved. As a result, the change detection of high resolution remote sensing imagery in coastal areas has become a focus in the research of remote sensing in recent years. Relevant research work is undoubtedly of great theoretical and practical significances.The main contents of this paper are:(1) By the analysis and summarizations of researches home and aboard, basic concepts about object-based change detection of remote sensing imagery are given. The "Pixel→Parcel→Change" model of change detection is proposed.(2) By repeated practice and research on more than one typical area, a complete system for image segmentation of coastal remote sensing imagery is built with combination of spectral and spatial geometry.(3) For the resolution of the decreased separability between the change class and the un-change class, and aboundant occlusions or shadows in coastal scenes, a parcel-based approach to change detection is proposed based on region covariance. This method can greatly improve the separability and reliability of changed class and unchanged class, as well as robust in different views and poses. (4) As complex scenes in coastal areas, a multi-scale parcel-based approach to change detection is proposed. With the help of multiscale analysis, this method can greatly improve the reliability of change detection.