基于Landsat-TM数据鄱阳湖湿地植被生物量遥感监测模型的建立

建立合理的经验模型对地面植被生物量进行遥感监测是可行的方法．利用遥感影像获得的植被指数和实地得到的生物量做回归统计分析，得到近似的地面生物量估算模型，这种模型的合理性取决于数据处理的水平和样本数据的大小．任何模型在具体应用时都需要根据实际情况进行改进，生物量估算模型更是如此．该文首先说明了回归分析的理论依据和适用条件；然后基于大量的实地测量数据，选取了若干种植被指数，分别进行了植被指数与生物量的线性和非线性回归分析，结果表明，非线性回归得到的结果更优．但是不同于以往的是，该文从统计学的角度对这种结果做了进一步说明，指出这种分析的局限性，说明了模型的适用条件，希望结果能为比较准确地实施遥感监测和湿地调查提供参考。

Establishment of wetland vegetation biomass model by in-situ and remote sensing observation in Poyang Lake area

To build reasonable remote sensing model of wetland biomass, the experimental model is a usually available method. Using vegetation index obtaining from TM satellite image and wetland vegetation bion/ass data from the corresponding field spots, the approximate simulating model is established on statistics analysis. The rationality of this kind of model depends on precision of image data processing and quantity of samples. Any model needs modification according to practical condition, especially for experimental model. Based on vast field investigation data of vegetation biomass, the remote sensing model of Xanthium grass and several other kinds of grasses was established, which is in a dominant community in area and quantity in Poyang Lake wetland area. The linear and non-linear regression analyses of Xanthium community and others are analyzed, and the result shows that non-linear regression analysis was more accurate. The conclusion need further interpreted by statistics method. The presented model methodology on the wetland biomass estimation by remote sensing is able to provide warranty for exploitation and protection of wetland vegetation resources.