论文部分内容阅读
基于决策树和EOS/MODIS数据,参考黑龙江省的农事活动和农作物发育期数据,对该省旱耕地进行了分类。结果表明,利用MODIS较高时间分辨率的特点,提取该省植被物候信息,可以在较短时间内实现对旱耕地分布及面积的定量描述;由于黑龙江省的干湿冷暖差异较大,按气候特征分区进行旱耕地判识的效果较好,总体精度可达到85.27%,Kappa系数0.8415。该项研究可为黑龙江省旱耕地面积数据的及时获取提供参考。
Based on the decision tree and EOS / MODIS data, the classified farmland in the province was classified with reference to the data of agricultural activities and crop development in Heilongjiang Province. The results show that using the characteristics of high temporal resolution of MODIS to extract the phenology information of the province’s vegetation can realize the quantitative description of the distribution and area of dryland in a short period of time. Because of the great difference between the warm and cold conditions in Heilongjiang Province, The characteristic zoning is better for dryland identification, with an overall accuracy of 85.27% and a Kappa coefficient of 0.8415. This study can provide a reference for the timely acquisition of dryland area data in Heilongjiang Province.