摘 要:冬小麦是我国重要的粮食作物之一,准确提取冬小麦种植区域范围对保证粮食安全具有重要意义。文章以山东省威海乳山市为研究区域,使用 GF-1C、GF-6、ZY-3 多源影像数据,提取乳山市冬小麦种植范围。根据乳山市冬小麦种植及生长情况,选取 4 月中旬至 5 月上旬最佳时期的卫星影像;用计算 NDVI 作为新波段替代红光波段与绿、蓝波段进行合成,更加突出植被信息;利用深度学习训练冬小麦提取模型,实现冬小麦种植范围的自动提取,提取精度为 94.39%,效果较好。
关键词:冬小麦;NDVI;深度学习;多源影像
DOI:10.19850/j.cnki.2096-4706.2023.02.029
中图分类号:TP18 文献标识码:A 文章编号:2096-4706(2023)02-0120-03
Research on Winter Wheat Extraction from Multi-Source Images Based on NDVI and Deep Learning
XUE Yu, YAO Jinming, LU Qinghui, YANG Ren
(Shandong Provincial Institute of Land Surveying and Mapping, Jinan 250013, China)
Abstract: Winter wheat is one of the important food crops in China. It is great significance to extract the range of planting area of winter wheat accurately for ensuring food security. In this paper, we take Rushan City, Weihai City, Shandong Province as the research area, and use GF-1C, GF-6, ZY-3 multi-source image data to extract the planting range of winter wheat in Rushan City. According to the planting and growth of winter wheat in Rushan city, satellite images of the best period from mid-April to early May are selected; calculate NDVI as a new band to replace the red band and the green and blue band to synthesize, more prominent vegetation information; deep learning is used to train the winter wheat extraction model to realize automatic extraction of the planting range of winter wheat with an extraction accuracy of 94.39%, it has a good effect.
Keywords: winter wheat; NDVI; deep learning; multi-source image
参考文献:
[1] 杨蕙宇,王征强,白建军,等 . 基于多特征提取与优选的冬小麦面积提取 [J].陕西师范大学学报:自然科学版,2020,48(1):40-49.
[2] 杨闫君,占玉林,田庆久,等 . 利用时序数据构建冬小麦识别矢量分析模型 [J]. 遥感信息,2016,31(5):53-59.
[3] 王碧晴,韩文泉,许驰 . 基于图像分割和 NDVI 时间序列曲线分类模型的冬小麦种植区域识别与提取 [J]. 国土资源遥感,2020,32(2):219-225.
[4] 王晓晓,韩留生,杨骥,等 .Sentinel-2 与 Landsat8 数据组合下的多特征冬小麦面积提取 [J]. 测绘通报,2022(3):111-115.
[5] 周亮,慕号伟,马海姣,等 . 基于卷积神经网络的中国北方冬小麦遥感估产 [J]. 农业工程学报,2019,35(15):119-128.
[6] 中国乳山网 . 乳山概况 [EB/OL].[2022-08-01]. http://www.rushan.gov.cn/col/col51333/index.html.
[7] 陈文志,许调娟,童英良 .GF-1B、C、D 星数据在国土资源调查监测领域的应用研究 [J]. 浙江国土资源,2022(3):40-43.
[8] 赵鸿飞,路钊,伊洋,等 . 基于 GF-6 的植被覆盖度遥感估测研究 [J]. 测绘与空间地理信息,2022,45(3):19-23.
[9] 郑冬梅,王海宾,夏朝宗,等 . 基于 ZY-3 卫星多光谱影像估算浙江省乔木林地上碳密度 [J]. 北京林业大学学报,2020,42(1):65-74.
作者简介:薛雨(1992—),女,汉族,山东济宁人,工程师,硕士,研究方向:遥感影像智能解译及地理信息数据处理。