基于Landsat8卫星数据的蝗虫遥感监测方法研究

区域性的蝗虫灾害对农业生产形成了巨大的危害。实现对蝗虫的发生发展的实时监测对于治蝗防蝗具有重要意义。以内蒙古赤峰市北部三旗为研究区域,通过对Landsat8 OLI卫星数据蝗虫寄主植物分类,结合叠加先验蝗区分布图,判别出蝗虫适生地。然后,采用Landsat8卫星数据,反演叶面积指数、地表温度和土壤湿度等蝗虫生境参数,并结合外业同步调查蝗虫生境数据、地表覆盖数据、历史蝗灾数据等进行分析与建模。同时,利用逐步回归分析得到了蝗虫虫口密度与叶面积指数、地表温度和土壤湿度等生境参数的关系模型,验证结果表明,监测模型具有较高的精度,决定系数为0.50,均方根误差为3.17。

Locust Remote Sensing Monitoring Methods Based on Landsat8 Satellite Data

Regional locust does great harm to agricultural production. Real-time monitoring of the development of locust is of great significance for the locust control. We took three counties in northern Chifeng City of Inner Mongolia as the study area. Firstly, we classified locust host plants by using the sophisticated remote sensing classification algorithm on the Landsat8 OLI data, overlapped with prior locust distribution regions, and distinguished the locust suitable bases regions. Then, we retrieved some important locust habitat parameters, such as leaf area index, land surface temperature and soil moisture by using Landsat8 satellite data. Meanwhile, the synchronous investigation data, land cover data, historical locust hazard data were combined for analysis and modeling. Finally, we used stepwise regression analysis to obtain the relationship between locust density and leaf area index, land surface temperature and soil moisture. The model results showed a high accuracy with R2 of 0.50 and RMSE of 3.17. It is indicated that the Landsat8 satellite data has a certain potential in locust remote sensing monitoring, and the research provides an important reference for similar studies.