摘要
水稻是中国的主要粮食作物,及时获取水稻种植面积和空间分布信息对指导水稻生产、调整区域供需平衡等具有重要的意义。以江苏省为例,利用2009—2011年连续三年的MODIS 8 d合成地表反射率数据(MODIS09A1),计算了归一化差值植被指数(normalized difference vegetation index,NDVI)、增强型植被指数(enhanced vegetation index,EVI)和陆表水指数(land surface water index,LSWI)。结合水稻在不同生长发育期EVI的时间序列变化特征,确定了水稻面积提取的关键生育期。根据水稻移栽期稻田土壤含水量高的特征,利用NDVI、EVI和LSWI三种指数构建判别条件,确定可能种植水稻的区域。利用线性光谱混合像元分解模型对包含水稻的混合像元进行分解,得到江苏省三年水稻种植空间分布。最后,选取研究区内的水稻典型样区,利用与MODIS同时期的较高分辨率的环境小卫星HJ-1 CCD(30 m)数据提取水稻种植面积和空间分布,以此作为参考数据进行精度验证,同时利用统计部门的江苏省水稻种植面积统计数据对江苏省水稻面积进行验证,两种方法验证后表明误差均在10%以内。研究表明,采用MODIS09A1数据结合线性光谱混合模型可以更高精度地提取大范围的水稻种植面积。
Paddy rice is one of the main crops in China.Timely information acquisition of rice planting area and spatial distribution at a large scale is of great significance in guiding rice production and regu- lating regional balance of supply and demand.In order to obtain paddy field area of Jiangsu Province, three types of vegetation indexes ( including NDVI ( normalized difference vegetation index), EVI ( en- hanced vegetation index), and LSWI (land surface water index)) are calculated by using the MO- DIS09A1 data from 2009 to 2011. Based on the temporal variation characteristics of EVI in different growth stages of rice ,the key growth period of rice area extraction extracts is determined.According to the characteristics of rice transplanting period with high soil moisture, NDVI, EVI and LSWI are used to identify potential planting area of rice. The linear spectral mixture model is applied to decompose the mixed pixel of potential rice area, and the rice spatial distribution is obtained in Jiangsu Province from 2009 to 2011.Finally, selecting the typical sample area of rice ,the rice planting area and spatial distribu-tion are extracted by using HJ-1 CCD(30 m) data,and the results are used as reference data to verify the extraction accuracy.Meanwhile, the statistical rice planting area of Jiangsu Province from the statisti- cal department is also used to verify the rice area.The error is within 10% by two methods of the statis- tical data and HJ-1 CCD(30 m) data.The research shows that it can more accurately extract a large range of rice planting area by using MODIS09A1 data and the linear spectral mixture model
出处
《大气科学学报》
CSCD
北大核心
2014年第1期119-126,共8页
Transactions of Atmospheric Sciences
基金
公益性行业(气象)科研专项(GYHY20090622)
江苏省"六大人才高峰"项目(NY-038)
教育部留学回国人员项目
江苏高校优势学科建设工程资助项目(PAPD)