基于多时相Landsat8数据的吉林梨树县玉米遥感估产研究

Remote Sensing Yield Estimation of Corn in Lishu County of Jilin Province Based on Multi-temporal Landsat8 Data

粮食产量关系到国计民生,世界各国都非常重视粮食估产在农业和粮食安全中的作用和意义,因此及时对吉林省重点玉米产区梨树县进行估产及动态预测粮食产量是非常必要的,也具有重要的现实意义。该文基于多时相Landsat8数据、全球高精度土地覆盖数据集GlobeLand30数据,结合吉林省统计年鉴数据对梨树县玉米种植区进行识别提取及精度检验,同时利用波段计算处理修正该区域植被指数,并以此建立较为准确的吉林梨树县玉米遥感估产模型。研究表明,累积NDVI值和累积RVI值均在玉米生长的抽雄期达到最值,且具有一定的负相关性;对于任何生长时期,使用累积NDVI值作为特征参数、对抽雄期及乳熟期进行估产的模型具有更高的精度数据。因此,多时相Landsat8数据在农作物遥感应用方面有较好的发展空间。

Grain output is related to the national economy and people's livelihood,and countries all over the world attach great importance to the role and significance of grain yield estimation in agriculture and food security.Therefore,it is very necessary to estimate and dynamically predict grain yield in Lishu County,a key corn producing area of Jilin Province.It also has important practical significance.Based on the multi-temporal Landsat8 data and the global high-precision land cover data set GlobeLand30 data,combined with the statistical yearbook data of Jilin Province,the corn planting area of Lishu County was identified and extracted and its accuracy was tested.At the same time,the regional vegetation index was modified by band calculation,and a more accurate remote sensing yield estimation model of corn in Lishu County of Jilin Province was established.The results showed that both the cumulative NDVI value and the cumulative RVI value reached the maximum at the heading stage of corn growth,and had a certain negative correlation;for any growth period,the model using cumulative NDVI value as characteristic parameter to estimate yield at heading stage and milk stage has higher accuracy data.Therefore,multi-temporal Landsat8 data has a good development space in the application of crop remote sensing.