资源科学 ›› 2019, Vol. 41 ›› Issue (3): 601-611.doi: 10.18402/resci.2019.03.17

• 生物资源 • 上一篇    

基于时序LAI的地块尺度玉米长势监测方法

苏伟1,2(), 朱德海1,2, 苏鸣宇3, 黄健熙1,2, 刘哲1,2, 郭浩1,2   

  1. 1. 中国农业大学土地科学与技术学院,北京 100083
    2. 农业部农业灾害遥感重点实验室,北京 100083
    3. 滕州市第一中学,滕州 277500
  • 收稿日期:2018-01-19 修回日期:2018-12-11 出版日期:2019-03-20 发布日期:2019-03-20
  • 作者简介:

    作者简介:苏伟,女,山东省滕州市人,博士,副教授,博士生导师,主要从事农业遥感应用研究。E-mail: suwei@cau.edu.cn

  • 基金资助:
    国家自然科学基金项目(41671433,41371327); 十三五国家重点研发计划项目(2017YFD0300903)

Field-scale corn growth monitoring using time series LAI

Wei SU1,2(), Dehai ZHU1,2, Mingyu SU, Jianxi HUANG1,2, Zhe LIU1,2, Hao GUO1,2   

  1. 1. College of Land Science and Technology, China Agricultural University, Beijing 100083, China
    2. Key Laboratory of Remote Sensing for Agri-Hazards, Ministry of Agriculture, Beijing 100083, China
    3. No.1 Middle School of Tengzhou, Tengzhou 277500, China
  • Received:2018-01-19 Revised:2018-12-11 Online:2019-03-20 Published:2019-03-20

摘要:

农作物长势监测可为田间管理措施调整和农作物产量预测提供及时、准确的信息。针对中国地块面积小的情况,本文采用中高分辨率的多源遥感影像Landsat-7 ETM+影像、Landsat-8 OLI影像、高分一号(GF-1)影像、HJ-1 A/B卫星影像为数据源;针对目前过多依赖NDVI等易饱和植被指数的问题,研究中选择LAI为长势监测指标,并利用PROSAIL辐射传输模型反演LAI,以保证该指标的准确性;长势实时监测采用RPLAI、LVCI、MLVCI指标,从多个角度表征长势的实时监测结果。研究结果表明:①同时相GF-1影像和Landsat-8 OLI影像具有较高的相关性,2种影像在近红外波段、绿波段、红波段的相关性决定系数R2分别为0.9320、0.7339、0.7153。②基于PROSAIL模型可以反演得到高精度的LAI,基于时序LAI的玉米生长过程监测结果表明:2015年,黑龙江农垦总局八五二农场6月下旬玉米冠层LAI快速增加,7月底、8月初LAI达到最大,并持续一段时间,进入8月下旬后,LAI开始下降。③利用RPLAI、LVCI、MLVCI指标对研究区玉米长势实时监测的结果表明,与2011—2014年相比,2015年八五二农场玉米长势一般,研究区北部长势较好,南部区域长势较差。从研究结果我们得出如下结论:①同时相的Landsat-8 OLI影像与GF-1遥感影像,经过相对辐射定标后可以结合使用于农作物长势监测中;②利用PROSAIL模型反演时序LA,可用于地块尺度的农作物长势精细监测。

关键词: 多源卫星影像, 地块尺度, LAI, PROSAIL模型, 生长过程监测, 实时监测

Abstract:

Crop growth monitoring is an important work for crop management and yield prediction. Aimed to solve the problem of mixed spectrum resulting from small agricultural land plots in China, this study used high and medium spatial resolution images including Landsat-7 ETM+, Landsat-8 OLI, GF-1, and HJ-1 A/B images for corn growth monitoring. In order to avoid too much reliance on the Normalized Difference Vegetation Index (NDVI) parameters, leaf area index (LAI) was selected as the corn growth monitoring parameter and the PROSAIL radiative transfer model was used to retrieve LAI. Three indices (LAI ratio to previous year (RPLAI), vegetation condition index based on LAI (LVCI), and mean vegetation condition index based on LAI (MLVCI)) were used for real-time monitoring of corn growth. The results of a case study in the 852 Farm of Heilongjiang Farms & Land Reclamation Administration in 2015 indicate that: (1) The simultaneous GF-1 image reflentance and the Landsat-8 OLI image reflentance are highly correlated. The correlation coefficients R2 of near-infrared bands, green bands, and red bands of the GF-1 image and the Landsat-8 OLI image are 0.9320, 0.7339, and 0.7153, respectively. This is the precondition for establishing the time series LAI for corn growth monitoring using multi-source remote sensing images. (2) The accuracy of retrieving LAI using PROSAIL radiative transfer model is high——the correlation coefficient R2 is 0.8030 and the root mean squared error (RMSE) is 0.7675. The retrieved time series LAI indicate that LAI increased quickly at the end of June, reached the maximum at the end of July or the beginning of August, and started to decrease at the end of August. (3) The RPLAI, LVCI, and MLVCI indices were used for the real-time monitoring of corn growth and the results indicate that the growth in 2015 was at an average level, and corn growth in the northern part of the area was better than in the southern part.

Key words: multi-source satellite images, field-scale, LAI, PROSAIL model, growth process monitoring, real-time monitoring