基于多时相遥感影像的大豆遥感提取  

Soybean Remote Sensing Extraction Based on Multiple Temporal Remote Sensing Images

作　　者：王珊 WANG Shan(Shandong Provincial Institute of Land Surveying and Mapping,Jinan 250013,China)

机构地区：[1]山东省国土测绘院,山东济南250013

出　　处：《测绘与空间地理信息》2024年第10期130-133,共4页Geomatics & Spatial Information Technology

摘　　要：准确掌握大豆种植面积对我国粮食安全具有重要的意义。本研究选择大豆生长周期内的全时相影像,开展大豆种植信息的高精度提取。通过本研究的开展,得到以下结论:通过多时相光谱对比,8月7日的大豆光谱区分度最高;通过多源指数计算与对比,筛选出了NDVI、DVI、NDWI、NDre1、NDre25种指数。通过面向对象分类,各地物均完整提取;提取的地块边界与实际的田埂走向吻合度较好,且增加多种指数的约束,农作物与其他植被的区分度较好,林地、草地、护坡灌草等对农作物提取的影响较小。并且通过后处理,较完整地提取出了所有大豆种植信息,对遥感提取的大豆种植面积进行提取精度验证,最终的遥感解译相关系数为0.7695,提取精度较高。It is of great significance to accurately grasp the soybean planting areas for food security in China.In this study,the full temporal images of soybean growth cycle were selected to carry out high-accuracy extraction of soybean planting information.Through the development of this study,the following conclusions were obtained:through multi-temporal spectral comparison,the degree of soybean spectral discrimination on August 7th was the highest;five indexes including NDVI,DVI,NDWI,NDre1 and NDre2,were selected through calculation and comparison of multi-source indexes.Through object-oriented classification,all objects are completely extracted;the extracted plot boundary is in good agreement with the actual field ridge trend,and a variety of index constraints are added.Crops are better differentiated from other vegetation,while forest land,grassland,slope protection shrubs and grasses have less influences on crop extraction.And through post-processing,more complete extraction of all soybean planting information is used to verify the extraction accuracy of soybean planting area extracted by remote sensing images.The final correlation coefficient of remote sensing interpretation is 0.7695 and the extraction accuracy is relatively high.

关 键 词：大豆 多时相 指数筛选 特征扩充 遥感提取 

分 类 号：P237[天文地球—摄影测量与遥感]