基于无人机遥感的农作物长势关键参数反演研究进展

无人机遥感是生态-环境-资源领域新兴的重要研究手段,近年来在农作物长势研究中得到了迅速的发展与应用。清楚、透彻地认识基于无人机遥感的农作物长势研究现状及存在问题,有利于更好地把握当前的核心领域并开展更进一步的研究。首先回顾了国际上"基于无人机遥感农作物长势研究"为主题的论文发表情况,其次对无人机遥感平台及不同传感器的基本遥感原理、反演的参数类型、各自的优势及局限性进行梳理,并概述了基于无人机遥感的农作物长势反演流程。在此基础上,一方面将农作物长势参数归纳为形态指标、生理生化指标、胁迫指标、产量指标等4类;另一方面将农作物长势参数的反演方法归纳为经验统计回归与机器学习法、形态特征与光谱特征识别法、辐射传输模型法、多角度航拍与卫星-无人机影像结合法等4类,并针对不同反演方法的优势与不足进行总结。最后综合国内外的研究现状进行了讨论分析与展望评价。该文通过综述近10 a来无人机遥感农作物长势关键参数反演的研究成果,可为今后基于无人机遥感方法的农作物长势研究的理论基础与技术支持方面提供参考。

Progress on key parameters inversion of crop growth based on unmanned aerial vehicle remote sensing

Unmanned aerial vehicle (UAV) remote sensing is an emerging and important research tool in the field of ecology, environment and resources. And it has been rapidly developed and applied in the study of crop growth in recent years. A clear and thorough understanding of the current situation and existing problems of crop growth based on UAV remote sensing is conducive to a better grasp of the current core areas and further research. In this study, we firstly reviewed the papers published on the topic of "research on crop growth based on remote sensing by unmanned aerial vehicle". Then, UAV remote sensing platform and the principle of basic remote sensing in different sensors, inversion parameter types, and its'' advantages and limitations were summarized. And the crop growth inversion process based on UAV remote sensing was summarized in detail. On this basis, crop growth parameters were summed up by divided into 4 categories, including morphological index, physiological and biochemical index, stress index, and yield index. On the other hand, the inversion methods of crop growth parameters were classified into 4 categories, including empirical statistical regression and machine learning, morphological features and spectral features identification, radiation transmission model, multi-angle aerial photography and satellite-UAV image combinations. And the advantages and disadvantages of different inversion methods were summarized. Finally, we conducted a discussion analysis and outlook evaluation based on the research status at domestic and abroad. This paper summarizes the research results of UAV remote sensing crop growth key parameters inversion in recent 10 years, which can provide reference for the theoretical basis and technical support of UAV remote sensing based crop growth research in the future. In this study, we think the reasons for the improvement of remote sensing inversion accuracy of UAV are as follows: First, the aerial photography height is low and close to the canopy of crops. The obtained information is less distorted and responsive, and it can reflect the small changes of crop phenotypes well. Second, the spatial scale of UAV remote sensing research is relatively small, which objectively not only excludes the heterogeneous factors that affect crop growth inversion at medium and large scales, but also can precisely control the variables required by the test purpose. And we believe that the real advantages of UAV remote sensing inversion of crop growth parameters are that: 1) UAV remote sensing is maneuverability and flexibility. 2) UAV remote sensing can provide higher resolution image data. 3) UAV remote sensing is sensitive to response of spatial heterogeneity information. To overcome the limitations of UAV remote sensing, our opinions are as follows: 1) Reducing equipment cost, improving flight duration and load capacity, improving flight stability, overcoming systemic and non-systematic errors to further improve the accuracy of inversion, are important direction for UAV remote sensing to be broken in the future. 2) Realizing platform docking between the UAV and the sensor as well as manipulating the same interface are the key technologies to be solved in the future. 3) Compared with satellite remote sensing, UAV remote sensing data exhibited some defects. The remote areas are not easily to reach for the UAV. Ideal data also can not be easily obtained as affected by various factors as scheduled. Cumbersome data preprocessing and so on. 4) UAV remote sensing in combined inversion of multi-type sensors is yet to be implemented, and the UAV remote sensing data acquisition on a long-time scale is implemented for different crop growth periods. On this basis, a staged and multi-parameter dynamic inversion model is constructed according to different physiological characteristics of crops and remote sensing information, which needs to be further explored. In summary, compared with traditional satellite remote sensing, UAV remote sensing has its unique advantages, but there are still some scientific problems to be further solved. However, the existing limitations of UAV remote sensing also fully illustrate that it has a vast prospect and great development potential in the study of crop growth.