A Reefs Extraction Method of Ocean Remote Sensing Image Based on Water Characteristics Index

摘要

Reefs and coastal zone are closely link of marine and land system. Research on the features and changes of the related ecological system and the surrounding environment has important significance. The environment of medium and small reefs and reef groups is uncommon, and it plays an important role in resource development and marine safety. The Worldview-II satellite remote sensing image is analyzed and processed. In view of the linear correlated redundant information of different bands, the principal component analysis transform is used to reduce the redundant data, and the effective band could be extracted. A reefs edge extraction algorithm based on normalization water characteristics index is proposed in this paper. Using this method, the radiation is greatly weakened for the non-water features, and enhanced the radiation differences between the seawater and other objects. In order to obtain the segmentation threshold, the maximum between-cluster variance algorithm is used. Through the edge superposition between reefs and seawater, the reefs edge is extracted from the Worldview-II satellite remote sensing image. As for the computational complexity and extraction precision, this method is better than traditional algorithms.