| 基于Sentinel-2多光谱数据和机器学习算法的冬小麦LAI遥感估算 |
| |
| 引用本文： | 史博太,常庆瑞,崔小涛,蒋丹垚,陈晓凯,王玉娜,黄 勇.基于Sentinel-2多光谱数据和机器学习算法的冬小麦LAI遥感估算[J].麦类作物学报,2021(6):752-761. |
| |
| 作者姓名： | 史博太 常庆瑞 崔小涛 蒋丹垚 陈晓凯 王玉娜 黄 勇 |
| |
| 作者单位： | (西北农林科技大学资源环境学院，陕西杨凌 712100) |
| |
| 基金项目： | 国家863计划项目(2013AA102401-2) |
| |
| 摘 要： | 为了丰富大田尺度下冬小麦叶面积指数的遥感估算方法并提高估算精度,以关中地区冬小麦为对象,基于Sentinel-2多光谱卫星数据与地面同步观测的冬小麦叶面积指数样点数据,应用偏最小二乘回归(PLSR)、反向传播神经网络(BPNN)和随机森林(RF)法构建冬小麦叶面积指数估算模型,进行区域冬小麦叶面积指数遥感反演。结果表明,Sentinel-2多光谱卫星影像中心842nm近红外B8波段与冬小麦叶面积指数相关性最好,样本总体相关系数为0.778;植被指数中反向差值植被指数(IDVI)与冬小麦叶面积指数相关性最好,样本总体相关系数为0.776。各种估算模型中LAI-RF模型预测效果最佳,r~2为0.72,RMSE为0.53,RE为16.83%。基于LAI-RF估算模型,应用Sentinel-2多光谱卫星数据较好地反演了研究区冬小麦叶面积指数区域分布,其结果总体上与地面真实情况接近,说明以Sentinel-2卫星影像数据建立LAI-RF估算模型,可应用于区域冬小麦LAI反演制图。
|
| 关 键 词： | Sentinel-2多光谱遥感 冬小麦 叶面积指数 偏最小二乘回归 反向传播神经网络 随机森林 |
LAI Estimation of Winter Wheat Based on Sentinel-2 Multis-Pectral Data and Machine Learning Algorithm |
| |
| SHI Botai,CHANG Qingrui,CUI Xiaotao,JIANG Danyao,CHEN Xiaokai,WANG Yun,HUANG Yong.LAI Estimation of Winter Wheat Based on Sentinel-2 Multis-Pectral Data and Machine Learning Algorithm[J].Journal of Triticeae Crops,2021(6):752-761. |
| |
| Authors: | SHI Botai CHANG Qingrui CUI Xiaotao JIANG Danyao CHEN Xiaokai WANG Yun HUANG Yong |
| |
| Abstract: | In order to enrich the method of remote sensing estimation of winter wheat leaf area index(LAI) at the field scale and improve the estimation accuracy,taking the winter wheat planting area in Guanzhong area as the research area, the single band reflectance and various vegetation indices were extracted by using Sentinel-2 multispectral satellite data combined with the wheat LAI data measured synchronously on the ground.Based on the optimal number of spectral variables, partial least squares regression(PLSR), back propagation neural network(BPNN) and random forest(RF) were used to build the winter wheat leaf area index estimation models.The results showed that: for single band reflectance, the correlation coefficient between the near-infrared B8 band with 842 nm center wavelength and 145 nm wave width is the best, and the overall correlation coefficient of samples is 0.778. For all kinds of vegetation indices, the correlation between IDVI and LAI of winter wheat is the best, and the overall correlation coefficient is 0.776.For the machine learning model, LAI-RF estimation model has the best prediction effect, with r of 0.72, RMSE of 0.53 and RE of 16.83%.Based on the LAI-RF estimation model, the Sentinel-2 multispectral satellite data was used to better invert the regional distribution of winter wheat LAI in the study area. It further illustrates that the LAI-RF estimation model established with Sentinel-2 satellite image data can be applied to regional winter wheat LAI inversion mapping. |
| |
| Keywords: | Sentinel-2 multispectral remote sensing Winter wheat LAI PLSR BPNN RF |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《麦类作物学报》浏览原始摘要信息 |
| 点击此处可从《麦类作物学报》下载免费的PDF全文 |
|
