| 摘要: |
| 目的 实现乡镇尺度准确的农业土地利用、作物分布快速制图。方法 文章选用Sentinel-2A遥感影像，基于时序遥感指数阈值分类、最大似然等方法完成各土地利用类型分类、作物空间信息提取，进行乡镇尺度农业土地利用制图，与基于WorldView-2融合图像的0.5 m分辨率农业空间信息图比较验证。结果 （1）研究区粮油作物用地、菜地、设施农用地和园地的面积精度良好，依次为92.93%、98.98%、95.71%和95.14%，与实际面积的差异在8%以内；（2）应用OSM道路数据和历史高分辨率影像等生产的地块边界、农田道路网、水渠网等空间信息与Sentinel-2A多光谱影像分类结果进行叠加制图，能提高10 m分辨率影像分类结果的地块边界信息精度。结论 使用多源信息融合方法和Sentinel-2A光学影像生产的乡镇农业土地利用、作物分布空间大数据，能有效提升乡镇尺度土地利用制图精度和时效性，为应用遥感大数据开展乡镇农业景观分析、农业产业结构优化调整等提供技术参考。 |
| 关键词: 多源遥感数据 土地利用 乡镇尺度 制图 |
| DOI：10.12105/j.issn.1672-0423.20210503 |
| 分类号: |
| 基金项目:现代农业学科建设推进工程项目“作物信息多源遥感智能提取及农业空间大数据挖掘研究”（2021XKJS077）；四川省科技计划项目“农业大数据资产化管理与分析应用系统”(2021YFG0028) |
|
| Mapping agricultural land use based on Sentinel-2A images |
|
Jiang Yi, Liu Zhongyou, Dong Xiuchun, Ren Guoye, Li Zongnan
|
|
Institute of Remote Sensing and Digital Agriculture，Sichuan Academy of Agricultural Sciences，Chengdu 610066，China
|
| Abstract: |
| Purpose The purpose of the study was to achieve accurate and rapid mapping of agricultural land use and crop distribution at the township scale.Method First of all，Sentinel-2A remote sensing images were selected to complete the classification of each land use type and crop spatial information extraction by various classification methods，such as time-series remote sensing index threshold classification，maximum likelihood method，etc. Then，the township-scale agricultural land use map was produced based on the above work. Finally，it was validated by comparing with the 0.5m resolution agricultural spatial information map based on WorldView-2 fused images.Result （1）The area accuracy of grain and oil crops land，vegetable land，agricultural land for facilities and garden land in the study area was good，with 92.93%，98.98%，95.71% and 95.14% in that order，and the difference with the actual area was within 8%；（2）The spatial information of plot boundaries，farmland road network and water channel network produced by applying OSM road data and historical high-resolution images was overlaid with Sentinel-2A multispectral image classification results for mapping，which can improve the accuracy of parcel boundary information in 10m resolution image classification results.Conclusion The spatial data of agricultural land use and crop distribution in townships produced by using multi-source information fusion method and Sentinel-2A optical images can effectively improve the accuracy and timeliness of township-scale land use mapping. It can provide technical reference for the application of remote sensing data to carry out rural agricultural landscape analysis，agricultural industry structure optimization and adjustment. |
| Key words: multi-source remote sensing data land use township scale mapping |
