基于物候特征参数的土地覆盖分类研究——以东北地区为例

Study on Land Cover Classification Based on Phenological Characteristic Parameters

基于1年MODIS旬最大值合成数据,运用Savizky-Golay滤波器对其进行了平滑滤波处理,应用Logistic模型提取了东北地区5个关键物候参数,并利用特征参数进行了研究区土地覆盖分类。结果表明,关键物候参数分类可明显提高植被覆盖类型的分类精度;其中,高盖度植被分类精度提高最为显著。

Based on panel data of the maximum MODIS values for each ten days,Savizky-Golay filter was adopted to deal with the data.Five key phenological parameters in the northeast areas were drawn with Logistic Model and land cover classification of research area was made with the characteristic parameters.Results showed that key phenological parameters classification could significantly improve the classification precision of vegetation cover types,in which,the increase of classification precision of high cover degree vegetation was the most significant.