基于RVoG模型的双极化SAR水稻株高反演  

Inversion of Rice Plant Height from Dual-Polarization SAR Data Based on RVoG Scattering Model

作　　者：付书娟 吴建 刘龙威 付海强[1] 朱建军[1] 李楠 宋晴 FU Shujuan;WU Jian;LIU Longwei;FU Haiqiang;ZHU Jianjun;LI Nan;SONG Qing(School of Geosciences and Info-physics,Central South University,Changsha 410083;Guangdong Provincial Land Resources Surveying and Mapping Institute,Guangzhou 510599;Key Laboratory for Monitoring of Tropical Subtropical Natural Resources in South China,Ministry of Natural Resources,Guangzhou 510663;School of Electronics and Communication Engineering,Sun Yat-Sen University,Guangzhou 510275,China)

机构地区：[1]中南大学地球科学与信息物理学院,湖南长沙410083 [2]广东省国土资源测绘院,广东广州510599 [3]自然资源部华南热带亚热带自然资源监测重点实验室,广东广州510663 [4]中山大学电子与通信工程学院,广东广州510275

出　　处：《地理与地理信息科学》2024年第4期28-33,共6页Geography and Geo-Information Science

基　　金：国家自然科学基金重大科研仪器研制项目(42227801)。

摘　　要：水稻株高被广泛应用于物候监测、水稻健康评估及产量预测等领域,极化合成孔径雷达(Polarimetric Synthetic Aperture Radar,PolSAR)信号能穿透水稻冠层记录水稻垂直结构信息,有助于实现高分辨率、大范围水稻株高提取。该文提出一种适用于非干涉条件下PolSAR数据水稻株高反演方法:利用极化SAR分解技术分离冠层及地表散射信号,引入随机地体二层散射模型(Random Volume over Ground,RVoG)对分解得到的水稻冠层散射能量进行建模,从而建立水稻株高反演模型;最后,联合邻域同质像素并利用NSGA-Ⅱ遗传算法对模型进行解算。利用2019年13景Sentinel-1影像在西班牙地区进行试验,水稻株高反演精度达到0.1 m,R^(2)达0.96以上,证明新方法能较好地适用于非干涉条件下PolSAR数据水稻株高反演。Rice plant height is widely used in the fields of climate monitoring,health assessment and yield prediction.Polarimetric Synthetic Aperture Radar(PolSAR)signals can penetrate the rice canopy and record the vertical structure information of rice,which provides the feasibility of high-resolution and large-scale rice plant height extraction.To realize rice plant height estimation,a correlation between PolSAR observations and rice biophysical parameters needs to be established.In view of this,this paper proposes a rice plant height inversion method for PolSAR data under non-interferometric conditions:first,the dual-polarization decomposition technique is used to separate the scattered signals from the canopy and ground surface;after that,the random volume over ground(RVoG)scattering model is introduced to model the scattered energy of rice canopy obtained from the decomposition,and the rice height inversion model is established;finally,the model is solved by combining neighborhood homogeneous pixels and NSGA-Ⅱgenetic algorithm.The experimental validation was carried out in the southern region of Seville in Spain using Sentinel-1 image in 2019,and the results showed that the accuracy of rice plant height inversion reached 0.1 m and R^(2) was above 0.96.It proves that the new method can be better applied to the rice plant height inversion of PolSAR data under non-interferometric conditions,and can provide strong support for realizing efficient rice plant height monitoring.

关 键 词：水稻株高反演 RVoG模型 双极化分解 

分 类 号：S127[农业科学—农业基础科学] S511