Coastline Extraction Using Support Vector Machine From Remote Sensing Image

Coastal zone is the transition zone which is from the ocean to the land. It is the important area of human development and utilization due to special geographical location and rich natural resources. Coastline is an important symbol of coastal zone. Meanwhile,it is regarded as a baseline for dividing the marine and terrestrial management areas. Different types of coastal landforms lead to the different types of shoreline. Therefore, there exist various interpretation keys and extracting methods for different shoreline types from remote sensing images. This paper briefly describes the impact on the environment due to changes in coastal zone. It also discusses the advantage of remote sensing application in coastal survey and dynamically monitoring etc. Moreover, the concept of shoreline in different researches is clarified and the interpreting keys of different coastline types are illustrated as well in this paper. The principles and methods of remote sensing imagery for coastline extraction are explained in detail. Taking the coastal area in Fujian province as an example, the characteristics of five typical coastline types in Landsat7 ETM+ image are analyzed and enhanced using different methods. For different coastal types I use different methods to analysis and enhance the image. After that, the SVM (support vector machine)method is adopted to classify the land cover in different coast area and the water-edges are detected. Finally, the modification of water-edges is fulfilled based on the coastline computation model, thereby the coastal line is completely extracted.