刘潭,许童羽,于丰华,袁青云,郭忠辉,王永刚.基于PROSAIL模型偏差补偿的水稻叶绿素含量遥感估测[J].农业机械学报,2020,51(5):156-164.
LIU Tan,XU Tongyu,YU Fenghua,YUAN Qingyun,GUO Zhonghui,WANG Yonggang.Remote Sensing Estimation of Rice Chlorophyll Content Based on PROSAIL Model Deviation Compensation[J].Transactions of the Chinese Society for Agricultural Machinery,2020,51(5):156-164.
摘要点击次数: 1495
全文下载次数: 746
基于PROSAIL模型偏差补偿的水稻叶绿素含量遥感估测   [下载全文]
Remote Sensing Estimation of Rice Chlorophyll Content Based on PROSAIL Model Deviation Compensation   [Download Pdf][in English]
投稿时间:2019-09-11  
DOI:10.6041/j.issn.1000-1298.2020.05.017
中文关键词:  水稻  叶绿素含量  光谱分析  PROSAIL模型偏差补偿  混合建模
基金项目:国家重点研发计划项目(2016YFD0200600)、辽宁省教育厅重点项目(LSNZD201605)、国家自然科学基金项目(61673281)、中国博士后科学基金项目(2018M631820)、辽宁省博士科研启动基金项目(2019-BS-207)和辽宁省自然科学基金指导计划项目(2019-ZD-0720)
作者单位
刘潭 沈阳农业大学 
许童羽 沈阳农业大学 
于丰华 沈阳农业大学 
袁青云 沈阳农业大学 
郭忠辉 沈阳农业大学 
王永刚 沈阳农业大学 
中文摘要:以东北水稻为研究对象,以提高叶绿素估测精度和模型可解释性为目标,提出了一种混合建模方法。以PROSAIL辐射传输机理模型为基础,模拟水稻冠层光谱,建立叶绿素含量的查找表,初步估测叶绿素含量,并采用最小二乘支持向量机(LSSVM)建立误差模型,对PROSAIL模型偏差进行补偿,弥补PROSAIL建模时产生的误差。为验证模型的估测能力,选取13种与作物叶绿素关系较为密切的植被指数,通过不同统计模型的模拟分析,筛选出4种较优的植被指数,分别建立单因子输入的最优预测模型(GNDVI、RSI、(SDr-SDb)/(SDr+SDb)的乘幂模型及MCARI的指数模型)。以4种植被指数作为输入,利用偏最小二乘法(PLS)、LSSVM回归法、BP神经网络及本文提出的混合建模方法分别构建水稻叶绿素含量多因子预测模型,并进行估测和验证。结果表明,相比单因子输入的最优预测模型,混合模型具有较低的预测偏差,其建模集R2=0.7406,RMSE为0.9852mg/dm2,验证集R2=0.7332,RMSE为1.0843mg/dm2。与采用其他多因子预测模型相比,本文方法具有较高的估测精度和良好的鲁棒性。另外,混合建模方法以PROSAIL模型为基础,物理意义较为明确,提高了预测模型的可解释性。本文可为作物叶绿素含量估测提供新的思路和方法,为诊断水稻氮营养含量和监测水稻长势提供参考。
LIU Tan  XU Tongyu  YU Fenghua  YUAN Qingyun  GUO Zhonghui  WANG Yonggang
Shenyang Agricultural University
Key Words:rice  chlorophyll content  spectral analysis  PROSAIL model bias compensation  hybrid modeling
Abstract:Accurate estimation of crop chlorophyll content using spectral information is an important part of field crop growth assessment and the basis for precise fertilization and scientific management of crops. The rice in Northeast China was taken as the research object, a new hybrid modeling method was proposed to improve the accuracy of chlorophyll estimation and model interpretability. Firstly, based on the PROSAIL model, the canopy spectra of rice was simulated, and a lookup table for chlorophyll content was established to initially inversion chlorophyll content. Then the least squares support vector machine (LSSVM) method was used to establish the error model to compensate the PROSAIL output deviation, which can compensate for the error caused by PROSAIL modeling. To verify the proposed model’s ability to estimate, totally 13 vegetation indices that were more closely related to crop chlorophyll was selected, and then the four optimal vegetation indices were screened out through the simulation analysis of different statistical models, and the optimal prediction model for single factor input was established, including power model for GNDVI, RSI, (SDr-SDb)/(SDr+SDb), exponent model for MCARI. In addition, combined with the four vegetation indexes as input, the multi-factor prediction model of rice chlorophyll content was constructed by using partial least square method (PLS), LSSVM, BP neural network and the proposed hybrid modeling method, and the predictive model was estimated and verified. The results showed that the hybrid model had a large advantage and a low prediction bias than the optimal prediction model with single factor input. The R2 of modeling set was 0.7406, the root mean square error (RMSE) was 0.9852mg/dm2;and the R2 of verification model was 0.7332, RMSE was 1.0843mg/dm2. Compared with other multi-factor prediction models, the proposed method also had certain advantages, with high estimation accuracy and good robustness. In addition, the hybrid modeling method was based on the PROSAIL model, which the physical meaning was clear and the interpretability of the prediction model was improved. Therefore, the proposed modeling method can provide ideas and methods for chlorophyll content inversion, and provide reference for the diagnosis of rice nitrogen and monitoring of rice growth.

Transactions of the Chinese Society for Agriculture Machinery (CSAM), in charged of China Association for Science and Technology (CAST), sponsored by CSAM and Chinese Academy of Agricultural Mechanization Science(CAAMS), started publication in 1957. It is the earliest interdisciplinary journal in Chinese which combines agricultural and engineering. It always closely grasps the development direction of agriculture engineering disciplines and the published papers represent the highest academic level of agriculture engineering in China. Currently, nearly 8,000 papers have been already published. There are around 3,000 papers contributed to the journal each year, but only around 600 of them will be accepted. Transactions of CSAM focuses on a wide range of agricultural machinery, irrigation, electronics, robotics, agro-products engineering, biological energy, agricultural structures and environment and more. Subjects in Transactions of the CSAM have been embodied by many internationally well-known index systems, such as: EI Compendex, CA, CSA, etc.

   下载PDF阅读器