基于多源数据的草本湿地边界提取方法

Method of Herbaceous Swamps’ Boundary Extraction Based on Object-Oriented

针对第二次湿地调查数据精度低难以满足草本湿地保护与再利用需求的问题,提出了利用高分二号影像提取草本湿地边界的面向对象决策树分类方法。首先,对Landsat 8影像进行目视解译,分析2015-01~2015-12Landsat 8影像草本湿地时相特征确定提取草本湿地边界的最佳时相;然后,对最佳时相的高分二号融合影像同其他辅助数据进行叠置分割,抽取湿度分量、归一化差分植被指数和色调特征用于决策树分类。结果表明,该方法提取的草本湿地边界与目视解译结果接近,比第二次湿地调查边界精度有显著提高。

Aiming at the problems of low precision of the second wetlands survey data,which is difficult to meet the demand of herbal wetlands protection and reuse,an object-oriented classification method based on decision tree is proposed to extract herbal wetlands’ boundary with high resolution remote sensing image.First of all,in order to choose the optimum phase,we visually interprete and analyze temporal characteristics of Landsat 8 images from January to December 2015.Then,overlap and apply image segmentation to GF2 image and assisted data.Extract humidity,NDVI,hue for decision tree classification.Results show that the boundary of the herbal wetlands extracted by this method is close to that of the visual interpretation,which is significantly more precise than that of the second wetlands survey boundary.