论文部分内容阅读
及时、准确监测水稻种植面积,对区域粮食政策制定、粮食安全以及农业发展具有重要意义。然而我国南方地区水稻生长期内降水充沛的气候特点使得遥感影像“云污染”现象严重,为解决水稻种植信息遥感提取存在可用数据不足的问题,以江汉平原为例,利用时空数据融合模型(Spatial and Temporal Data Fusion Approach,STDFA)将Landsat 8 OLI与时序MODIS数据融合,重构出具有高时-空分辨率特征数据,然后采用面向对象的SVM分类方法对研究区内水稻种植信息进行提取,结果如下:融合后的红与近红外波段反射率与真实反射率的相关系数分别为0.84和0.81,研究区水稻提取精度为94.46%,Kappa系数为0.91。说明时空融合模型能够较好地重构出高时空分辨率数据,从而实现多云雨地区农作物种植信息遥感提取。
Timely and accurate monitoring of rice acreage is of great importance to the formulation of regional food policies, food security and agricultural development. However, the climatic characteristics of abundant precipitation during the rice growing season in the southern part of China make the phenomenon of “cloud pollution” serious. In order to solve the problem of lack of available data for remote sensing extraction of rice plantation information, taking the Jianghan Plain as an example, (Spatial and Temporal Data Fusion Approach, STDFA), the Landsat 8 OLI and temporal MODIS data were fused to reconstruct the feature data with high spatio-temporal resolution. Then the object-oriented SVM classification method was used to extract rice plantation information in the study area The results are as follows: The correlation coefficients between reflectivity and true reflectivity in the red and near infrared bands after fusion were 0.84 and 0.81, respectively. The extraction precision of rice in the study area was 94.46% and the Kappa coefficient was 0.91. It shows that the spatio-temporal fusion model can reconstruct the high spatial and temporal resolution data well, so as to realize remote sensing extraction of crop planting information in the cloudy rain area.