基于物候特征参数的橡胶树种植信息遥感提取研究  

Research on Remote Sensing Extraction of Planting Information for Rubber Trees Based on Phenological Characteristic Parameters

作　　者：田光辉[1] 李海亮[2] 陈汇林[1] 

机构地区：[1]海南省气象科学研究所/南海气象减灾防灾重点实验室,海口570203 [2]中国热带农业科学院科技信息研究所/海南省热带作物信息技术应用研究重点实验室,海南儋州571737

出　　处：《中国农学通报》2013年第28期46-52,共7页Chinese Agricultural Science Bulletin

基　　金：国家自然科学基金项目"基于时空耦合和数据同化的作物单产过程摸拟研究"(40971218);农业部农业信息技术重点实验室开放基金"天然橡胶干旱动态监测及其影响定量评估"(2012006)

摘　　要：为了更好地监测橡胶树的生长状态,客观准确地获取橡胶树的种植的地理位置信息,采用非对称高斯函数对海南岛内MODIS EVI时间序列数据进行滤波处理,定义作物物候特征参数并计算海南岛每个像元的特征参数值;统计分析基于MODIS EVI数据的海南岛橡胶树样区作物关键物候特征参数值变化规律;根据橡胶树样区特征参数值变化规律建立海南岛橡胶树种植信息提取算法,提取结果采用不同比例的橡胶树混合像元分类法表示。根据提取算法对海南岛2011年橡胶树种植信息进行提取,用验证数据对该提取算法进行精度检验与实用性评价,检验结果显示该算法的平均分类精度达85.97%,可以用来提取橡胶树种植信息。In order to better monitor the growth state of rubber trees, and objectively and accurately get geographic location information of rubber tree planting, asymmetric Gaussian function was used to filter MODIS EVI time series data on Hainan Island, phenological characteristic parameters were defined, and characteristic parameters of each pixel was calculated; change rules for key phenological characteristic parameters of crops in sample region of rubber trees on Hainan Island were summarized and analyzed based on MODIS EVI data;a planting information extraction algorithm for rubber trees on Hainan Island was established based on aforesaid characteristic parameters, and extraction results were expressed by mixed pixel classification method of different proportions. Planting information for rubber trees in 2011 was extracted based on the extraction algorithm, and accuracy test and practicability evaluation to the algorithm were conducted using validation data. Test result showed that the algorithm could reach an average classification accuracy of 85.97%, thus applicable for extracting planting information of rubber trees.

关 键 词：高斯函数 橡胶树 信息提取 海南岛 

分 类 号：S1[农业科学—农业基础科学]