基于高光谱影像的渤海海冰遥感监测研究

Study on Bohai seaice monitoring based on hyperspectral remote sensing imagery

监督分类方法是海冰遥感监测中常用的有效方法,但不同的监督分类方法以及波段选择,在海冰识别中的精度有较大差异。为提高海冰监测的精度,本文使用高光谱传感器提供的可见光、近红外波段的连续成像光谱信息,对比了不同的波段组合在多种分类方法中海冰提取的精度,分析了不同波段组合、不同分类方法在海冰监测上的优缺点,最终得出海冰监测中最佳的波段组合以及最适宜的监督分类方法。

As an effective method,supervised classification is widely used in seaice remote sensing monitoring field,but different supervised classification methods and bands combinations always result in different classification results and accuracy.In order to improve the precision of the seaice monitoring,this paper used hyperspectral remote sensed imagery,which provides the visible light,near-infrared wavelengths of continuous imaging spectra data,compared seaice extracting accuracy of the different band combinations,analyzed the advantages and disadvantages of seaice monitoring using different bands combinations and different classification methods,and proposed the best bands combination and the most suitable supervised classification method for the seaice monitoring.