Research for extraction of aquaculture water information in coastal zone using CBERS-02B data based on object-oriented technique

In recent years, remote sensing technology has been widespread applied in coastal resources survey and ecological environment research. The methods of coastal object extraction have been steadily innovating. Compared with traditional manual interpretation and classification based on pixel, the object-oriented technique can increase classification accuracy and reduce the salt - pepper noise considerably. With the increase of the spatial reference of remote sensing and more textural feature, the object-oriented technique has broad application prospects. In this article, object-oriented method is applied to extract aquaculture water using CBERS-02B image. The method includes the following: image segmentation, segment merging, spectral and spatial features analyses, other water body elimination. The result shows that the method is effective in accurately extracting aquaculture water information by using CBERS - 02B data. © 2010 IEEE