结合实测的盐渍地表光谱数据,采用经验线性法对遥感影像进行大气校正。分析了实测高光谱数据与遥感影像光谱值的关系,然后针对盐生植被光谱的相似性,对影像进行纹理特征提取。结合影像的纹理特征和光谱特征,运用神经网络分类法进行盐渍区地表覆被分类。结果表明辅以纹理特征的分类方法有效提高了盐生植被的分类精度。 更多还原
【Abstract】 The paper firstly analyzed the relation between the measured spectral data and multi-spectral image data after atmospheric correction by empirical line calibration method.Then textural features were extracted using the multi-spectral ALOS image.Classification and recognition were carried out by the neural network method combined with the spectral characteristics and spatial characteristics on the synthetic image.The classification accuracy was improved. 更多还原