基于谱-时特征分析的农作物重金属胁迫遥感识别方法研究

Crop spectral variability under various types of environmental stress is similar, how to apply remote sensing technique to monitor heavy-metal stressed crops has been an unresolved scientific issue. Heavy metal stress in crops is characterized by stability in time, which differs from other stressors that are typically more transient (e.g., water, pests/diseases, and nutrient stress). Thus, this research explores the principles and methods of monitoring heavy metal contamination in crops based on the time characteristics analysis for parameters response to environmental stress. Here, heavy metal pollution area cultivated rice in Zhuzhou, Hunan province is selected as the study area. Firstly, Analysis of crop parameters response to environmental stresses, namely, the changes of appearance, color and physiological function of crop under environmental stress and its spectral diagnostic parameters. Secondly, on the basis of crop parameters response to environmental stresses in the whole-growth stage of multi-year, crop parameters response to environmental stresses are simulated by using the improved crop growth model and remote sensing data. Using multi-scale wavelet analysis, the characteristics information of heavy metal-stressed rice is diagnosed and extracted. Furthermore, the mechanism analysis and statistical methods are used to establish the remote sensing dynamic model of rice canopy under heavy metal pollution stress. The purpose of this study is to reveal the constant and stable pattern of the crop under heavy metal contamination, and build the remote sensing calculation model of rice canopy under heavy metal pollution stress. This study will provide scientific basis for monitoring heavy metal contamination on rice using remote sensing, and provide new ideas for monitoring various environmental stress using remote sensing technology.

由于农作物受各种环境胁迫所引起的光谱变异特征具有相似性，因此，尽管运用遥感监测农作物重金属胁迫的研究比较多，却未能得到很好的解决。本项目从重金属胁迫与其它环境胁迫效应所呈现的时间特征差异入手，对农作物重金属胁迫遥感监测的关键科学问题展开研究。选取湖南株洲重金属污染区为实验区，分析环境胁迫下水稻胁迫特征变化及其诊断光谱参数；运用谱-时特征分析方法，通过遥感与改进的作物生长模型同化获取水稻环境胁迫响应参数在生长期及年际之间的时间变化特征，采用小波变换等时间序列分解方法对参数变化特征进行分解，提取水稻重金属胁迫特征信息；依据机理分析与统计方法，建立水稻重金属胁迫遥感识别模型。通过本项目的研究，揭示水稻重金属胁迫谱-时特征变化规律，建立水稻重金属胁迫遥感模型，为各类环境胁迫遥感监测提供研究思路。

如何精确地识别农作物重金属污染胁迫对于人类健康、食品安全具有重要意义。本项目选取了典型重金属污染区的“自然农田生态系统”作为实验区，针对农作物重金属胁迫遥感探测所面临的关键科学问题，开展了以下工作：构建了农作物重金属胁迫精确识别的评估指标，探索了农作物胁迫响应特征的谱-时分析方法，构建农作物重金属胁迫水平评估模型，并探究了大尺度农作物重金属污染胁迫遥感动态监测的高效计算方法。主要的研究成果包括：根据农作物受到环境胁迫下的光谱响应，借助遥感影像与物理模型，构建了农作物重金属胁迫响应的敏感光谱参数，实现了水稻重金属胁迫水平的精确判别；为了探索重金属胁迫与其它环境胁迫效应所呈现的时间特征差异性，运用时间序列曲线分解算法，实现了作物重金属胁迫稳定信息的提取,建立了胁迫敏感光谱指数的时间变异系数；基于非参数模型算法构建的融合光谱信息和时空信息的重金属胁迫水平计算模型，有效提高了作物重金属胁迫的甄别精度；面向区域大尺度水稻重金属胁迫动态监测的需要，探索了农作物参数时空变化的高效计算策略。.本课题的研究对水稻重金属污染胁迫遥感监测所面临的科学问题进行了较为系统的探究，探讨了“自然农田生态系统”农作物重金属污染识别的机理与方法，建立了农作物重金属胁迫遥感识别模型，实现了水稻重金属污染胁迫准确识别，促进了农作物重金属污染胁迫遥感监测技术应用的可行性，为自然农田生态系统农田重金属遥感监测提供了参考。