深度学习驱动的土壤信息感知技术:进展、挑战与展望

李佳怿 ,  刘楠 ,  杨玮 ,  李民赞

中国农业大学学报 ›› 2025, Vol. 30 ›› Issue (10) : 73 -90.

PDF (2869KB)
中国农业大学学报 ›› 2025, Vol. 30 ›› Issue (10) : 73 -90. DOI: 10.11841/j.issn.1007-4333.2025.10.07

深度学习驱动的土壤信息感知技术:进展、挑战与展望

作者信息 +

Deep learning‑driven soil information sensing: Advances, challenges, and prospects

Author information +
文章历史 +
PDF (2937K)

摘要

农田土壤信息的快速感知是智慧农业的重要支撑,为厘清土壤信息感知技术发展中存在的问题,对深度学习驱动下土壤信息感知的研究现状进行系统梳理与分析。重点聚焦不同感知尺度和感知方法下,各类深度学习模型在破解传统土壤信息反演难题中的创新实践,深入分析各类方法与思路的优势与局限性。结果表明:1)对于近端感知系统,深度学习通过构建各类深度神经网络模型,有效解决了传统电性能传感器校准中环境因素耦合困难、适应性差的问题,提升了土壤水分、盐分及养分原位检测的精度与鲁棒性;同时,深度学习方法在解决可见-近红外光谱预测有机质、全氮含量时面临的水分干扰、跨时空泛化性差等关键问题上展现出显著优势。2)各类深度学习模型与方法广泛应用于处理星载和机载遥感数据,构建了更精准的土壤有机碳、水分、盐分等参数的大范围反演模型,显著提升了遥感数据的重建、校正与降尺度能力。3)深度学习驱动的多模态、多尺度与多时相信息融合模型方法得到广泛应用;各类轻量级网络、混合模型以及注意力机制的应用,实现了各类异构数据的有效交互与融合,显著增强了土壤有机质、水分、全氮等多参数同步预测的精度和实时性。4)从各项研究趋势中发现,发展适应复杂农田环境、具备强泛化能力的反演模型,以及构建基于群智感知与协同感知的近端-机载-星载多尺度土壤信息感知网络是重要趋势;同时,结合土壤连续体模型理论,未来需深入探索土壤微生物菌群动态与有机质周转过程的原位监测方法与定量关联模型,以实现从单一理化参数监测向土壤生态系统健康评估的跨越。

Abstract

The rapid sensing of farmland soil information is as a improtant support for smart agriculture. To clarify the existing issues in the development of soil information sensing technologies, a systematic review and analysis are conducted on the current research status of soil information sensing driven by deep learning. This study focuses is placed on the innovative practices of various deep learning models in addressing traditional challenges associated with soil information inversion under different sensing scales and methods, along with an in-depth examination of the strengths and limitations of these methods and approaches. The findings indicate: 1) For proximal sensing systems, deep learning has effectively addressed the coupling complexity of environmental factors and poor adaptability in traditional electrical sensor calibration by constructing various deep neural network models, thereby enhancing the accuracy and robustness of in-situ detection for soil moisture, salinity, and nutrients. Meanwhile, deep learning methods demonstrate significant advantages in overcoming key challenges such as moisture interference and poor cross-temporal/spatial generalization in Vis-NIR spectroscopy-based prediction of soil organic matter/total nitrogen content. 2) A wide range of deep learning models and methods have been applied to process spaceborne and airborne remote sensing data, enabling the development of more accurate large-scale inversion models for soil organic carbon, moisture, salinity, and other parameters. These approaches have substantially improved the reconstruction, calibration, and downscaling capabilities of remote sensing data. 3) Deep learning-driven multimodal, multi-scale, and multi-temporal information fusion models have gained widespread adoption. The integration of lightweight networks, hybrid models, and attention mechanisms has facilitated effective interaction and fusion of heterogeneous data, significantly enhancing the accuracy and real-time performance of synchronous multi-parameter prediction. 4)Emerging research trends highlight the importance of developing inversion models with strong generalization capabilities for complex agricultural environments and constructing multi-scale (proximal-airborne-spaceborne) soil information sensing networks based on swarm intelligence and collaborative sensing. Furthermore, integrating soil continuum model theory, future efforts should prioritize the exploration of in-situ monitoring methods and quantitative correlation models for soil microbial dynamics and organic matter turnover processes, aiming to transition from single physicochemical parameter monitoring to comprehensive soil ecosystem health assessment.

Graphical abstract

关键词

土壤养分 / 深度学习 / 传感器 / 智慧农业

Key words

soil nutrients / deep learning / sensors / smart agriculture

引用本文

引用格式 ▾
李佳怿,刘楠,杨玮,李民赞. 深度学习驱动的土壤信息感知技术:进展、挑战与展望[J]. 中国农业大学学报, 2025, 30(10): 73-90 DOI:10.11841/j.issn.1007-4333.2025.10.07

登录浏览全文

4963

注册一个新账户 忘记密码

土壤是农业生产的核心载体,其功能价值贯穿于作物生长的全生命周期。作为农业生产体系的物质基石,土壤不仅为植物根系构建了三维支撑框架,更通过复杂的理化过程持续供给作物生长所需的水分、矿质营养及微量元素。在田间管理实践中,土壤含水率、电导率、土壤有机质(Soil organic matter, SOM)与土壤全氮(Soil total nitrogen,STN)等构成评估土壤质量的关键指标体系[1-2],其中,SOM通过胶结作用促进团粒结构形成,显著改善土壤孔隙度与通气性,其分解产物可持续释放氮、磷、钾等营养元素及中微量元素,形成动态平衡的养分供应库。自然生态系统中土壤全氮含量与有机质储量存在显著正相关关系[3-5]。但现代农业体系下,土壤养分与土壤水盐参数拥有复杂的耦合关系,同时受包括如轮作、连作模式的种植制度差异、灌溉策略、有机肥施用强度等多重人为干预因素影响[6]。快速、精确地获取与反演土壤信息对农业生产的发展具有重要的意义。
传统土壤分析多采用化学分析方法。土壤有机质的实验室检测方法主要包括以灼烧法为代表的质量损失法和以Walkley-Black氧化法为代表的化学测定方法[7],如干燃法[8]和重铬酸钾氧化-分光光度法[9];土壤全氮检测主要采用凯氏定氮法和杜马斯燃烧法[10-11],或是使用元素分析质谱仪[12];土壤水分、盐分的测定则通常分别依赖于烘干法和浸提液电导率测量。这些检测方法精度高、可靠性强,但涉及复杂的样本处理流程,需要较高时间、人力及耗材成本,无法快速高效地获取土壤信息,难以满足精细农业和智慧农业的效率要求。随着光学、电化学传感技术的发展,快速、无损的土壤信息感测方法及技术在过去25年内取得了快速发展。如在砂土和壤土中,土壤电导率值随着含水量的增大而线性增大,基于介电常数或电导率传感器可以较好地关联土壤水盐信息[13-14];土壤水分、有机质与全氮中的羟基(-OH)、硝基(-NO₂)、碳碳双键(C=C)、氨基(-NH₂)等官能团在红外光谱中具有特定的伸缩振动吸收峰,通过光学透射、漫反射信息对其分频、泛频波段的吸收系数进行观测,可以进一步实现土壤有机质、全氮含量的反演。然而,土壤传感、遥感方案还同时面临着许多挑战,如传感器测得土壤的EC值不仅与土壤中的有机质、K含量高度相关[13],同时受到温度、土壤孔隙、植物根系等环境因素的影响;基于漫反射遥感光学传感器存在多种物质的光谱信息存在解混困难,抗扰能力差等问题[15],这些难题在大范围的田间实践中仍具有较大挑战。传统方法的高成本低效率与现有传感技术的环境敏感性双重瓶颈,严重限制了传统土壤信息感知与反演方法的鲁棒性与泛化能力。
21世纪10年代后,以卷积神经网络 (Convolutional neural network,CNN)、Transformer为代表的深度学习模型与方法开启了人工智能连接主义(Artificial intelligence connectionism)的新阶段[16]。连接主义认为智能行为源于大量简单单元(如神经元、卷积模块)通过动态连接关系实现的集体行为,而非预设的符号规则,如人工神经网络(Artificial neural network,ANN)、CNN,通过梯度下降调整连接权重实现知识表示和学习[17-18]。连接主义的人工智能兴起对传统基于人工推导、测试、预设、调制的土壤感测设计、研究方法与模型构建形成了冲击,为解决土壤原位信息变异性强、环境复杂难题提供了新视角。如深度学习极大地拓宽了反演模型的信息容量,单一模型可以构建多种土壤的信息反演模型为田间泛化提供了执行基础[19-20];此外基于深度学习特征对齐、特征融合等前沿方法为波谱、电化学等多源异构数据融合提供了新理论[21-23];更多的,深度学习强化了土壤科学、光谱学等领域间交叉协同,进一步促进了土壤信息感知技术的发展。本研究拟从土壤养分感知系统与土壤信息反演模型2个角度,梳理分析人工智能及其相关技术驱动下土壤感知技术的研究进展,展望未来发展趋势,以期为构建更高效、精准的土壤信息感知体系提供理论参考和技术指导。

1 土壤信息近端感知系统

目前,基于遥感、传感技术的土壤感知方法及技术形成了较为成熟的研究方法和技术手段,根据信息采集平台的工作尺度,可主要分为星载遥感、机载遥感、传感与近端遥感。其中,传感与近端遥感手段是精细获取土壤参数的关键,根据采集信息的不同可主要分为电性能、力学、气体及波谱手段。本研究基于文献收集,分析整理得到不同土壤反演参数和应用场景下近端土壤感知的相关技术手段见图1。

1.1 深度学习与土壤水分感知

针对土壤含水率(Soil moisture content, SMC)的传感器可主要分为体积含水率传感器和土壤基质势(Soil matric potential)传感器[24-25],前者主要通过间接方法关联土壤体积含水率,如以时域反射计(Time domain reflectometer,TDR)、电容探针为代表的介电技术,后者则主要通过测量从土壤中提取水分所需的力,利用土壤水分特征曲线反推土壤的体积含水率,如土壤张力计[26]。基于电磁技术的传感器因其成本优势,在田间测试的可重复性、及其在校正、应用中的便利性和高效性而广受欢迎[27],然而土壤的电磁特性同时受到多种因素的耦合(表1),传统的线性校准方法和模型,难以同时满足不同时空场景下土壤水分的最佳反演效果[28-29],进一步导致目前不同研究之间的校准方法缺乏标准化和一致性。

针对传统线性与非线性校准方法的适应性差,外部环境参数耦合难度高,需现场校准频繁等问题,越来越多的研究开始应用深度学习对电性能传感器的校准方法与反演模型进行优化[49-50]。如Palta等[51]基于5层深度的深度神经网络(图2),结合0.2~14 GHz频段的电磁信息与包含土壤质地、含水率与温度的多变量试验数据,解决了土壤介电行为与频率、温度、含水量及质地的非线性耦合难题,该模型实现了对介电常数实部和虚部超高精度测量,平方根误差RMSE(Root mean squared error)分别低至0.025 8和0.033 6,决定系数R²达0.999 8,显著优于响应面法等传统方法。

石玉娇等[52]通过构建BP神经网络模型,揭示了尿素质量与土壤水分传感器输出值的周期性振荡关系,解决了施肥干扰导致的测量误差问题,实现了误差峰值仅2.07%、平均绝对误差为0.64%的精准预测。Jafaryahy等[53]利用前馈神经网络整合土壤多频段电导率与介电特性,解决了同时精准预测土壤钾含量(误差<0.05 g/kg)和水分(误差<0.4%)的技术难题。Aranda Britez等[54]基于土壤饱和点和田间持水量2个特征点结合深度学习方法,提出了一种土壤水分传感器自校准方法,解决了传统校准方法成本高、不实用的问题,实现了在无需大量地面参考数据的情况下,快速、准确地校准土壤湿度传感器,显著提高了土壤湿度测量的准确性。相较于传统人为标定与校准体积含水率传感器方法,深度学习的校准方法通过学习多种环境参数与土壤含水率之间的复杂关系,动态地影响测量值或反演值的输出,让体积含水率传感器在多区域场景下拥有更好的反演效果。

1.2 深度学习与SOM、STN感知

对于土壤有机质(SOM)和土壤全氮(STN)含量等更具有反演难度的土壤参数,除了使用介电参数和土壤表观电导率,其他感测技术也得到进一步研究和应用,如可见-近红外光谱、拉曼光谱、人工嗅觉、便携式X射线诱导荧光(Portable X-ray fluorescence)和激光诱导击穿光谱(Laser-induced breakdown spectroscopy, LIBS)等[55-56]。经典的SOM预测模型构建方法常利用优化算法筛选关键波段进一步结合机器学习算法实现某一时空场景下精确预测。但不同区域、土壤类型背景下筛选的SOM关键波段存在差异(表2),受制于成土母质与土壤环境的高变异性,这类建模无法形成统一的标准,在跨时空迁移学习中具有较大的挑战。其中,土壤湿度、土壤颗粒等因素还与SOM、STN等参数之间的光谱信息存在复杂的非线性耦合,这类实验室建模方法往往还需要将土壤进行烘干、研磨、过筛等处理,以保证关联光谱信息的精度。这些关键问题严重阻碍了SOM、STN等参数原位信息的大范围、多时相的精细反演。

为克服使用可见-近红外光谱预测土壤参数时环境因素影响,国内外研究人员采用了各种方法,例如直接标准化(Direct standardization, DS)、分段直接标准化(Piecewise direct standardization, PDS)、外部参数正交化(External parameter orthogonalization, EPO)、正交信号校正(Orthogonal signal correction, OSC)和光谱空间变换(Spectral space transformation, SST)[73-75]。Ji等[76]使用DS算法对样本田间光谱进行转换,并结合局部加权回归模型进行SOM、STN的含量预测,预测决定系数R²分别达到0.641和0.658。Cao等[77]在研究内蒙古草原土壤时发现,EPO算法在不同水分梯度条件下对低土壤有机碳(Soil organic carbon,SOC)样本表现出更强的适应性,构建的EPO-RF模型对SOC的预测效果最优。Roudier等[78]通过分析不同水分梯度下的新西兰土壤样本,比较EPO和DS消除水分干扰的效果,研究结果发现当使用大量的样本用于水分校正时,EPO和DS的效果相似,但使用小样本时,EPO的消除效果优于DS。Ji等[79]进一步采用PDS方法对田间光谱进行校准,相较于直接利用原始光谱数据,SOC预测决定系数R²由0.03提高至0.71,并且所需校准样本数量少于DS方法。杨斌[80]利用辐射传输模型对野外条件采集的光谱进行水分消除,然后对SOM含量进行预测,水分校正之后模型的预测决定系数R2和相对分析误差分别提升了59.61%和67.13%。然而,传统校正方法如DS、PDS、EPO、OSC与SST等均属于线性校正范畴,其精度高度依赖于特定时空背景下的土壤特性。这类方法在应对跨时空场景的土壤信息反演时存在显著局限性:以EPO校正为例,其通过强化SOM预测能力需构建干扰因素投影矩阵,但该过程要求预先采集相同SOM含量、不同含水率的土壤波谱数据。当反演区域发生时空迁移时,必须重新采集数据并重构投影矩阵,导致方法可移植性受限。

随着深度学习的发展和应用,基于数据驱动的参数解耦或校正方法或基于跨时空海量数据训练的大型深度学习模型进一步得到重视。这类方法依托深度神经网络的非线性拟合能力与高容量特征提取优势,有望突破传统线性校正方法适应性差的瓶颈,为复杂时空场景下的土壤信息精准反演提供全新解决方案。如Zhao等[81]开发了一种基于深度学习的MIRNet模型,实现了水分干扰效应的非线性生成,有效抑制水分对可见/近红外光谱中有机碳预测的干扰,该模型构架见图3,采集到的土壤原始光谱数据会同时输入到上下文提取模块和光谱提取模块,提取原始光谱的光谱特征和上下文信息。然后,通过连接层将提取的参数进行结合,拼接后的数据会输入到频谱重建模块,从而得到只包含水分信息的光谱。将原始光谱数据与水分相关的光谱数据相减,得到估计的干土光谱。通过计算测量的干土光谱和估计的干土光谱之间的D-S损失(Distance-shape loss),调整模型的反向传播参数。模型预测决定系数R²达到0.926,整体性能优于EPO与OSC等传统校正方法。

即使使用线性校正方法,深度学习仍有强化模型泛化效果的潜力,Bao等[82]提出了一种结合Kubelka-Munk 辐射传输模型水分消除策略(Moisture elimination strategy,MSE)和混合深度学习模型(Multimodal transformer mechanism-CNN-convolutional long short-term memory,MT-CNN-ConvLSTM)的模型新范式,依靠LSTM架构提取光谱特征、ConvLSTM模块提取多尺度空间特征,再利用自注意力机制对融合特征进行学习,最终在湿润和半湿润地区均表现出优异的预测性能(R²≈0.76~0.77),显著提升了模型的空间迁移能力。在模型的构建思路上,部分学者利用深度学习的强大拟合能力,跳过校正步骤,直接对多时空背景下的海量土壤数据进行学习训练。如Singh等 [83]利用长短期记忆网络(LSTMs)结合主成分分析的方法和技术,利用深度学习自动捕捉土壤理化参数与光谱特征间的复杂映射关系,解决了从欧盟光谱库中的高光谱数据中量化土壤性质的问题,并在有机碳等多种土壤参数中实现了优异的预测效果,显著优于传统模型的效果。这些案例表明,相较于依赖先验假设的传统方法,依靠深度学习模型的高容量特征提取能力与强大的非线性拟合效果,可有效缓解跨区域土壤光谱变异性强、背景干扰因素复杂等问题。

深度学习模型也被用于融合多源土壤数据变相校正单一数据中存在的期望误差,实现更高的反演稳定性。如Gallios等[84]利用低成本近红外扫描仪和地理空间协变量,通过双输入深度学习模型,解决了光谱传感器预测在大范围区域中预测性能不佳的问题,实现了表层SOC预测准确性的显著提升,平方根误差RMSE降低了30.56%。Hong等[85]通过融合可见-近红外与近红外光谱数据,采用并行输入-卷积神经网络(Parallel input CNN)模型解决了单一传感器预测SOC的稳定性问题,在多区域土壤样本验证中实现了更精准的SOC预测效果。可以发现,得益于深度学习对数据耦合的强力辅助,土壤参数反演方法呈现多样化的趋势,并不局限于特定的研究思路与方法,既可利用深度学习方法用于强化土壤信息校正效果,也可基于深度学习模型利用足够样本丰度和数据源实现更好的反演稳定性。

基于便携式X射线诱导荧光、拉曼光谱、人工嗅觉等技术的土壤参数感知研究也具有相同的发展趋势。Yang等[86]通过结合基于迭代自适应窗口经验小波变换和高斯卷积的信息预处理方法与分层深度神经网络技术,解决了土壤重金属定量分析中的背景干扰与多元素同步检测难题。Liu等[87]采用基于热裂解-人工嗅觉系统的多特征优化方法,结合蒙特卡罗交叉验证异常样本剔除和遗传算法优化的反向传播神经网络特征降维技术,实现了土壤总氮含量更精准的定量预测,显著提升了人工嗅觉系统在土壤养分检测中的可靠性。

2 土壤信息空天遥感感知系统

星载和机载平台也是土壤信息监测的关键环节,星载平台根据采集的遥感波段可主要分为高光谱与多光谱成像(0.4~3 μm)、热红外遥感(3~1 000 μm)、和合成孔径雷达(Synthetic aperture radar,SAR)(5~1 000 mm)平台[88-89]。而基于机载平台则多为高光谱、多光谱、RGB成像平台、激光雷达等传感器[90]。表3示出2015—2025年不同深度学习模型在星载、机载平台的部分应用案例。利用星载、机载遥感数据实现土壤有机碳的精确反演具有较高的挑战性,不难发现,近年来研究人员基于深度学习技术通过庞大的土壤光谱数据库和联动近端遥感设备为高空遥感平台提供精准的波谱特征参考,以构建更精准的估算模型[91-94]。

除了用于构建更加精细的反演模型,深度学习技术还被进一步用于辅助遥感数据的重建和校正。已有较多研究利用深度学习方法对离散数据进行复杂拟合,形成连续高精度的时空数据,弥补传感器在时空分辨率的不足,Wei等[108]提出了一种物理约束的插值方法(PhyFill),该方法通过结合深度学习中的部分卷积神经网络技术和物理过程(如降水事件和土壤湿度干涸过程)来填充土壤湿度数据中的缺失部分,从而生成全球范围内连续且高精度的每日土壤湿度数据。Huang 等[109]提出了一种融合多源遥感数据的深度学习框架,通过结合降尺度辅助变量和偏差校正方法,生成了连续空间的土壤水分数据,该方法在奥地利HOAL传感器网络上的验证表明,其预测精度显著优于传统方法。Wei等[110]提出了一种基于时空信息的残差密集网络(Spatiotemporal residual dense network,SMDN),通过建模低分辨率地球科学参数与土壤湿度的关系,并结合残差校正模块,生成了高精度的土壤湿度地图。此外,利用深度学习强大的特征挖掘能力,在实现遥感数据高精度压缩和降噪具有较高的可行性,如Zhao等[111]通过构建基于深度强化学习的特征选择框架并结合多子集评估策略,利用马尔可夫决策过程建模与强化学习代理优化,有效解决了高光谱遥感图像在土壤有机物预测中的信息冗余与噪声干扰问题,预测精度与效率得到了显著提高。

3 土壤多模态、多尺度与多时相的感测信息融合

基于多源信息融合以提升反演效果一直是土壤信息反演方法的重要研究方向[112-113]。如美国Veris公司开发的iScan土壤传感装备同时测量土壤可见光反射率、红外反射率、土壤电导率等,拥有较好的土壤信息反演稳定性[114]。Bonecke等[115]通过结合非商业化的Geophilus测量系统和商业化的Veris MSP平台采集土壤伽马辐射活动、pH、表观电阻率的数据生成高分辨率土壤酸化地图。Wang等[116]基于电流-电压六端法与近红外光谱技术,通过建立土壤电导率与光谱反射率的指数关系及敏感波长校正模型,开发了车载式土壤参数检测系统,实现了土壤电导率与含水量的高精度在线测量。然而传统的基于多源信息融合的土壤信息反演模型多依赖于特征对齐的思路,需要人工设计数据的对齐、配准流程,让数据在张量维度、分辨率上达到较高的一致性。Meng等[91]提出了一种土壤有机碳含量预测模型,通过结合2009—2019年无云裸土期多时相的Landsat TM/OLI数据和2019年无云裸土期高光谱数据,配合包含坡度、海拔、坡向、曲率和地形高差的数字高程模型进行几何配准,进一步基于区域加权离散小波变换(Regional weighted discrete wavelet transform,RW-DWT),对时相-空间-光谱信息(Temporal-spatial-spectral,TTS)进行多维融合,其方法构架见图4。

综合比较可见,基于特征对齐思路传统方法往往需要人为设计复杂繁琐的流程,与后续的模型结构存在割裂,这一问题使得传统融合方法在应用中存在数据压缩过程的信息丢失与数据配准过程中的误差累加问题。

与数据合并、加权融合、小波变化融合等传统的数据融合方法相比,深度学习系统性地推进了多模态的数据交互和融合,特别是具有空间、通道、时相维度的多源异构数据:图像(二维空间和光谱通道)、点云信息(三维空间)、多时相信息(包含时序维度的图像、点云)等信息。依靠深度学习的自动微分与参数优化器,精细地实现各类数据的特征提取与压缩融合,可以避免关键特征的缺失和对齐误差,这些优势进一步影响了农情信息感知方法的设计与构建[117-118]。Li等[119]通过融合特征波长、可见光图像和热成像图像的多模态数据,结合轻量级多模态网络(MobileNet-LSTM和FCN分支网络),解决了单传感器信息有限及田间土壤水分变化干扰的问题,实现了土壤有机质和水分含量的高精度实时预测,其网络架构见图5,整体模型由2个分支网络和1个特征融合模块组成。第一个分支网络融合RGB图像和热成像数据,利用Mobile Net V3-LSTM模块进行高级特征映射;第二个分支网络处理光谱特征波长数据,通过全连接层进行非线性变化处理。最后,在特征融合模块中,采用交互注意力机制将2个分支网络提取到的特征张量进行融合。融合后的特征通过三层全连接层进行处理,最终实现对土壤水分、有机质和全氮含量的预测。

对于旋翼无人机平台,Yin等[100]通过包含RGB、多谱系和热成像的多模态图像数据结合深度学习技术,准确估算了苜蓿根际的土壤湿度,其中深度神经网络模型在多模态数据融合和土壤湿度估计方面表现最佳,并在不同灌溉条件下表现稳定。相似地,Wu等[120]设计了多模态无人机遥感平台(包括RGB、红外线和多光谱数据)结合卷积神经网络与长短期记忆网络的混合模型(CNN-LSTM),实现了对柑橘果园不同深度土壤水分的快速反演。Jiang等[103]通过结合高光谱和雷达遥感数据,利用深度学习网络构建了土壤光谱校正策略,有效减少了土壤表面水分和粗糙度对土壤有机碳估算的耦合效应,显著提高了遥感图像对土壤光谱的预测精度和地图能力。针对多时相的时空光谱数据,Rukhovich等[121]利用多时相Landsat遥感数据与U-Net卷积神经网络,结合光谱邻域土壤线技术和地理信息系统,解决了传统土壤退化检测方法耗时且难以在大范围内应用的问题,实现了裸土表面的自动识别和土壤退化地图的高精度构建,其中退化分布图的准确率达到了90%。利用深度学习模型对多维张量的强大支持性,可以有效将频谱、图像等多源异构数据进行特征对齐和特征融合,进一步提高反演的稳定性与精度。

4 未来与展望

当下研究呈现出适应性强、多参数耦合的土壤参数反演模型的建模趋势,以及多尺度、多模态的信息交互与融合的遥感、传感趋势。本文综述了人工智能驱动下的大田土壤信息感知技术,特别是在土壤参数反演模型建模以及遥感、传感领域多尺度、多模态信息交互与融合方面的最新进展。通过回顾与分析,可以明确了2个核心发展趋势:一是基于深度学习技术实现适应性强、多参数耦合的土壤参数反演模型的趋势;二是多尺度、多模态的信息交互与融合的传感系统设计。在土壤参数反演模型建模方面,随着深度学习技术的兴起,传统的基于人工推导和预设规则的方法逐渐被数据驱动、模型自适应的方法所取代。这些新型模型不仅能够处理复杂的非线性关系,而且能有效融合多源异构数据,提高反演精度和泛化能力,形成连续精细的时空反演数据。联动多尺度、多单位、多平台的农情协作感知(Cooperative sensing)与群智感知(Crowd sensing)将会是未来土壤参数农情感知系统发展的重要思路。

随着土壤参数获取精度和便利性的提升,如何利用丰富的土壤数据量化土壤菌群与不同农艺间的作用关系可能是一个关键研究方向。传统的观点认为土壤有机质(SOM)的持久性主要归因于某些难分解的有机碳形式[122]。然而,这些“难以分解的腐殖质大分子”在现代分析技术中并未得到证实,因此其存在性越来越受到质疑。针对这一问题,Lehmann等[123]提出了土壤连续体模型(Soil continuum model,SCM),认为土壤有机质是一个从完整植物材料到高度氧化的羧酸中“碳的连续体”。这一模型强调分解者生物对土壤有机质的可接近性,以及土壤矿物质对有机质分解的保护作用。Lehmann等[124]认为传统上通过增加“稳定腐殖质”池来提高土壤有机质的策略可能是徒劳的,因为土壤有机质是一个动态分解的连续体。因此,管理土壤有机质的关键在于促进其周转而非积累。大田土壤中的微生物活性和种类丰富度显著影响着土壤有机质的积累与循环[125],不同于森林土壤的SOM构成,草地与农田的SOM来源是真菌和细菌,二者占比总和均超过了55%,远高于森林土壤的33.4%[126]。由微土壤菌群加工的低分子量化合与矿物表面结合而受保护而进一步形成的矿物结合态有机质(Mineral-associated organic matter, MAOM),拥有更长的持久性,是农田、牧场等低碳土壤的主要SOM来源[127]。但由于种植和耕作的影响,农田MAOM表现出较高的脆弱性,长期的耕作与化肥的过度施加会导致农田土壤菌群的失衡与MAOM的快速流失,进而造成土壤的退化与板结。因此,高质量农田的建设仅仅关注当前农田的水分、SOM含量和全氮等元素类指标可能存在不足,土壤菌落健康程度的监测方法和其与SOM的时空联系在未来应当得到学者的重视和研究。

5 结 论

1)近端土壤感知系统通过深度学习技术显著提升了土壤参数反演的精度与适应性。基于电性能、光谱等多源传感数据,深度神经网络有效解决了传统校准方法中环境因素耦合困难、现场校准频繁等问题,实现了介电常数、土壤水分及养分的高精度预测。自校准方法与多频段数据融合技术进一步降低了传感器成本,推动了原位土壤信息的高效获取。

2)空天遥感土壤感知系统依托星载/机载平台的高光谱、多光谱及SAR数据,结合深度学习模型显著提升了土壤有机碳、水分、盐分等参数的大范围反演能力。深度学习不仅优化了光谱校正与信息融合流程,还通过物理约束插值、残差网络等技术实现了遥感数据的缺失填补与高精度重建,为全球土壤湿度监测提供了连续、高分辨率的数据支持。

3)多模态、多尺度与多时相信息融合技术通过深度学习系统性整合了光谱、图像、点云及时序数据,有效解决了单传感器信息有限及环境干扰问题。各类轻量级网络与混合模型实现了可见光、热成像、光谱数据的交互融合,显著提升了土壤有机质、水分等参数的实时预测精度。多时相遥感数据与深度学习的结合进一步增强了模型的时空泛化能力。

4)在模型方法层面,发展适应复杂农田环境的强泛化能力反演模型,通过多源数据融合与跨尺度学习提升模型对时空变异的适应性,同时探索群智感知与协同感知技术,构建覆盖近端-机载-星载多尺度的土壤信息感知网络;展望未来,随着海量感测数据的采集与应用,在机理认知层面,结合土壤连续体理论,需加强土壤微生物菌群动态与有机质周转过程的关联研究,开发针对MAOM的原位监测技术,并探索菌群结构、农艺措施与土壤健康指标的定量关系模型,以实现从单一理化参数监测向土壤生态系统功能评估的跨越。

参考文献

[1]

Karlen D L, Veum K S, Sudduth K A, Obrycki J F, Nunes M R. Soil health assessment: Past accomplishments, current activities, and future opportunities [J]. Soil and Tillage Research, 2019, 195: 104365

[2]

Sorenson P T, Shirtliffe S, Bedard-Haughn A K. Examining the effect of soil organic carbon on major Canadian Prairie crop yields with predictive soil mapping[J]. Agronomy Journal, 2024, 116(6): 2976-2989

[3]

童珊, 曹广超, 张卓, 刁二龙, 闫欣. 土壤全碳全氮空间异质性及影响因素分析:以祁连山南坡黑河上游为例[J]. 土壤, 2022, 54(5): 1064-1072

[4]

Tong S, Cao G C, Zhang Z, Diao E L, Yan X. Analysis of spatial heterogeneity and influencing factors of soil total carbon and nitrogen: Take the upper reaches of Heihe River on southern slope of Qilian Mountain as an example[J]. Soils, 2022, 54(5): 1064-1072 (in Chinese)

[5]

Prout J M, Shepherd K D, McGrath S P, Kirk G J D, Haefele S M. What is a good level of soil organic matter? An index based on organic carbon to clay ratio [J]. European Journal of Soil Science, 2021, 72(6): 2493-2503

[6]

Nelson D W, Sommers L E. Total carbon, organic carbon, and organic matter [M]. In: Methods of Soil Analysis. Bognor Regis: John Wiley & Sons, Ltd, 1996: 961-1010

[7]

Huang T T, Yang N, Lu C, Qin X L, Siddique K H M. Soil organic carbon, total nitrogen, available nutrients, and yield under different straw returning methods [J]. Soil and Tillage Research, 2021, 214: 105171

[8]

Miyazawa M, Pavan M A, de Oliveira E L, Ionashiro M, Silva A K. Gravimetric determination of soil organic matter[J]. Brazilian Archives of Biology and Technology, 2000, 43(5): 475-478

[9]

Allison L E, Bollen W B, Moodie C D. Total carbon[C]. In: Methods of Soil Analysis. Bognor Regis: John Wiley & Sons, Ltd, 1965: 1346-1366

[10]

Walkley A, Black I A. An examination of the degtjareff method for determining soil organic matter, and a proposed modification of the chromic acid titration method[J]. Soil Science, 1934, 37(1): 29-38

[11]

陈春霏, 付洁, 梁鹏, 孙媛. 土壤中全氮分析方法的研究进展[J]. 中国环境监测, 2018, 34(1): 112-119

[12]

Chen C F, Fu J, Liang P, Sun Y. Researche progress on analytical methods for total nitrogen in soil[J]. Environmental Monitoring in China, 2018, 34(1): 112-119 (in Chinese)

[13]

秦琳, 黄世群, 仲伶俐, 周虹, 赵珊, 向冰, 雷绍荣. 杜马斯燃烧法和凯氏定氮法在土壤全氮检测中的比较研究[J]. 中国土壤与肥料, 2020(4): 258-265

[14]

Qin L, Huang S Q, Zhong L L, Zhou H, Zhao S, Xiang B, Lei S R. Comparison of Dumas combustion and Kjeldahl methods for determining total nitrogen content in soil[J]. Soil and Fertilizer Sciences in China, 2020(4): 258-265 (in Chinese)

[15]

GB/T 42490—2023土壤与生物样品中有机碳含量与碳同位素比值、全氮含量与氮同位素比值的测定 稳定同位素比值质谱法 [S].北京:中国标准出版社, 2023

[16]

GB/T 42490—2023,carbon isotope ratio,total nitrogen content,and nitrogen isotope ratio in soil and biological samples-stable isotope ratio mass spectrometry Determination of organic carbon content [S]. Beijing: China Standard Press, 2023 (in Chinese)

[17]

Fragkos A, Loukatos D, Kargas G, Arvanitis K G. Response of the TEROS 12 soil moisture sensor under different soils and variable electrical conductivity[J]. Sensors, 2024, 24(7): 2206

[18]

Louki I I, Al-Omran A M. Calibration of soil moisture sensors (ECH2O-5TE) in hot and saline soils with new empirical equation [J]. Agronomy, 2023, 13(1): 51

[19]

Xie A Y, Zhou Q W, Fu L, Zhan L C, Wu W H. From lab to field: Advancements and applications of on-the-go soil sensors for real-time monitoring [J]. Eurasian Soil Science, 2024, 57(10): 1730-1745

[20]

Zhang B, Zhu J, Su H. Toward the third generation artificial intelligence[J]. Science China Information Sciences, 2023, 66(2): 121101

[21]

Alzubaidi L, Zhang J L, Humaidi A J, Al-Dujaili A, Duan Y, Al-Shamma O, Santamaría J, Fadhel M A, Al-Amidie M, Farhan L. Review of deep learning: Concepts, CNN architectures, challenges, applications, future directions [J]. Journal of Big Data, 2021, 8(1): 53

[22]

Puri M, Solanki A, Padawer T, Tipparaju S M, Moreno W A, Pathak Y. Chapter 1 - introduction to artificial neural network (ANN) as a predictive tool for drug design, discovery, delivery, and disposition: Basic concepts and modeling [C]. In: Artificial Neural Network for Drug Design, Delivery and Disposition. Boston: Academic Press, 2016: 3-13

[23]

Kok M, Sarjant S, Verweij S, Vaessen S F C, Ros G H. On-site soil analysis: A novel approach combining NIR spectroscopy, remote sensing and deep learning [J]. Geoderma, 2024, 446: 116903

[24]

Escorcia-Gutierrez J, Gamarra M, Soto-Diaz R, Pérez M, Madera N, Mansour R F. Intelligent agricultural modelling of soil nutrients and pH classification using ensemble deep learning techniques[J]. Agriculture, 2022, 12(7): 977

[25]

Wang Q, Liu Y F, Xiong Z T, Yuan Y. Hybrid feature aligned network for salient object detection in optical remote sensing imagery[J]. IEEE Transactions on Geoscience and Remote Sensing, 2022, 60: 5624915

[26]

Tang W, He F Z, Liu Y. TCCFusion: An infrared and visible image fusion method based on transformer and cross correlation[J]. Pattern Recognition, 2023, 137: 109295

[27]

Li J Y, Zeng H Y, Huang C X, Wu L B, Ma J, Zhou B B, Ye D P, Weng H Y. Noninvasive detection of salt stress in cotton seedlings by combining multicolor fluorescence–multispectral reflectance imaging with EfficientNet-OB2 [J]. Plant Phenomics, 2023, 5: 125

[28]

Mane S, Das N, Singh G, Cosh M, Dong Y. Advancements in dielectric soil moisture sensor Calibration: A comprehensive review of methods and techniques[J]. Computers and Electronics in Agriculture, 2024, 218: 108686

[29]

Bobrov P P, Belyaeva T A, Kroshka E S, Rodionova O V. Soil moisture measurement by the dielectric method[J]. Eurasian Soil Science, 2019, 52(7): 822-833

[30]

Dragonetti G, Farzamian M, Basile A, Monteiro Santos F, Coppola A. In situ estimation of soil hydraulic and hydrodispersive properties by inversion of electromagnetic induction measurements and soil hydrological modeling[J]. Hydrology and Earth System Sciences, 2022, 26(19): 5119-5136

[31]

Soulis K X, Elmaloglou S, Dercas N. Investigating the effects of soil moisture sensors positioning and accuracy on soil moisture based drip irrigation scheduling systems[J]. Agricultural Water Management, 2015, 148: 258-268

[32]

Stadler A, Rudolph S, Kupisch M, Langensiepen M, van der Kruk J, Ewert F. Quantifying the effects of soil variability on crop growth using apparent soil electrical conductivity measurements[J]. European Journal of Agronomy, 2015, 64: 8-20

[33]

Topp G C, Davis J L, Annan A P. Electromagnetic determination of soil water content: Measurements in coaxial transmission lines[J]. Water Resources Research, 1980, 16(3): 574-582

[34]

Wensink W A. Dielectric properties of wet soils in the frequency range 1–3000 MHz1 [J]. Geophysical Prospecting, 1993, 41(6): 671-696

[35]

邵芸,吕远,董庆,韩春明. 含水含盐土壤的微波介电特性分析研究[J]. 遥感学报, 2002, 6(6): 416-423

[36]

Shao Y, Lu Y, Dong Q, Han C M. Study on soil microwave dielectric characteristic as salinity and water content[J]. Journal of Remote Sensing, 2002, 6(6): 416-423 (in Chinese)

[37]

Mensah C, Katanda Y, Krishnapillai M, Cheema M, Galagedara L. Multifrequency electromagnetic induction soil moisture characterization under different land uses in western Newfoundland[J]. Canadian Journal of Soil Science, 2023, 103(3): 446-461

[38]

Savin I V, Muzalevskiy K V, Mironov V L. A dielectric model of thawed and frozen Arctic organic soils at 435 MHz[J]. Remote Sensing Letters, 2022, 13(5): 452-459

[39]

Mironov V L, Kosolapova L G, Savin I V. Dielectric model of thawed and frozen organic soil at the AMSR radiometer frequency[J]. Izvestiya, Atmospheric and Oceanic Physics, 2021, 57(12): 1783-1788

[40]

Kim H N, Park J H. Monitoring of soil EC for the prediction of soil nutrient regime under different soil water and organic matter contents[J]. Applied Biological Chemistry, 2024, 67(1): 1

[41]

Malicki M A, Plagge R, Roth C H. Improving the calibration of dielectric TDR soil moisture determination taking into account the solid soil[J]. European Journal of Soil Science, 1996, 47(3): 357-366

[42]

Cai J C, Wei W, Hu X Y, Wood D A. Electrical conductivity models in saturated porous media: A review[J]. Earth-Science Reviews, 2017, 171: 419-433

[43]

Friedman S P, Robinson D A. Particle shape characterization using angle of repose measurements for predicting the effective permittivity and electrical conductivity of saturated granular media[J]. Water Resources Research, 2002, 38(11): 18-1-18-11

[44]

Datsios Z G, Mikropoulos P N. Characterization of the frequency dependence of the electrical properties of sandy soil with variable grain size and water content[J]. IEEE Transactions on Dielectrics and Electrical Insulation, 2019, 26(3): 904-912

[45]

Heil K, Schmidhalter U. Characterisation of soil texture variability using the apparent soil electrical conductivity at a highly variable site[J]. Computers & Geosciences, 2012, 39: 98-110

[46]

周林虎, 王昊宇, 张秉来, 祁兆鑫, 曹荣泰, 范延彬, 孙生海, 刘宇平. 硫酸盐渍土表观电导率与水分、盐分及粒径关系研究[J]. 干旱区研究, 2021, 38(4): 1020-1030

[47]

Zhou L H, Wang H Y, Zhang B L, Qi Z X, Cao R T, Fan Y B, Sun S H, Liu Y P. The relationship between ECa of sulfate saline soil and moisture content, salt content, and particle size [J]. Arid Zone Research, 2021, 38(4): 1020-1030 (in Chinese)

[48]

Xiong W C, Shao Y. Models for imaginary part of dielectric constant of moisture and salt soil[C]. In: IGARSS 2004.2004 IEEE International Geoscience and Remote Sensing Symposium. Anchorage: IEEE, 2004: 4302-4305

[49]

Gao L, Song X N, Leng P, Ma J W, Zhu X M, Hu R H, Wang Y F, Zhang Y N, Yin D W. Impact of soil salinity on soil dielectric constant and soil moisture retrieval from active microwave remote sensing[J]. IEEE Transactions on Geoscience and Remote Sensing, 2022, 60: 4514112

[50]

Rentz J W, Sri Ranjan R, Ferguson I J, Holländer H M. Effects of salinity and water content on apparent conductivity in an alluvial setting in the Canadian Prairies[J]. Environmental Earth Sciences, 2021, 80(3): 93

[51]

张贤龙, 张飞, 张海威, 李哲, 海清, 陈丽华. 基于光谱变换的高光谱指数土壤盐分反演模型优选[J]. 农业工程学报, 2018, 34(1): 110-117

[52]

Zhang X L, Zhang F, Zhang H W, Li Z, Hai Q, Chen L H. Optimization of soil salt inversion model based on spectral transformation from hyperspectral index[J]. Transactions of the Chinese Society of Agricultural Engineering, 2018, 34(1): 110-117 (in Chinese)

[53]

Wu S Y, Zhao T J, Pan J M, Xue H Z, Zhao L, Shi J C. Improvement in modeling soil dielectric properties during freeze-thaw transitions[J]. IEEE Geoscience and Remote Sensing Letters, 2022, 19: 2001005

[54]

Brevik E C, Fenton T E, Horton R. Effect of daily soil temperature fluctuations on soil electrical conductivity as measured with the geonics® EM-38[J]. Precision Agriculture, 2004, 5(2): 145-152

[55]

Seyfried M S, Grant L E, Du E, Humes K. Dielectric loss and calibration of the Hydra probe soil water sensor[J]. Vadose Zone Journal, 2005, 4(4): 1070-1079

[56]

Zanetti S S, Cecílio R A, Silva V H, Alves E G. General calibration of TDR to assess the moisture of tropical soils using artificial neural networks[J]. Journal of Hydrology, 2015, 530: 657-666

[57]

Singh G, Panda R K, Bisht D S. Improved generalized calibration of an impedance probe for soil moisture measurement at regional scale using Bayesian neural network and soil physical properties[J]. Journal of Hydrologic Engineering, 2021, 26(3): 04020068

[58]

Palta P, Kaur P, Mann K S. Dielectric behavior of soil as a function of frequency, temperature, moisture content and soil texture: A deep neural networks based regression model [J]. Journal of Microwave Power and Electromagnetic Energy, 2022, 56(3): 145-167

[59]

石玉娇, 尹峥, 王红叶, 刘晓辰, 石庆兰, 梅树立. 尿素对土壤水分传感器测量精度的影响[J]. 农业机械学报, 2020, 51(S2): 388-394, 407

[60]

Shi Y J, Yin Z, Wang H Y, Liu X C, Shi Q L, Mei S L. Effect of urea on measurement accuracy of soil moisture sensor[J]. Transactions of the Chinese Society for Agricultural Machinery, 2020, 51(S2): 388-394, 407 (in Chinese)

[61]

Jafaryahya J, Keshavarz R, Abolhasan M, Lipman J, Shariati N. Integrating electrical features for simultaneous prediction of soil moisture and potassium levels based on neural network prediction model[J]. IEEE Transactions on Instrumentation and Measurement, 2025, 74: 2513912

[62]

Aranda Britez D, Tapia A, Millán Gata P. A self-calibration algorithm for soil moisture sensors using deep learning[J]. Applied Intelligence, 2025, 55(4): 276

[63]

Rosin N A, Demattê J A M, Almeida Leite M C, de Carvalho H W P, Costa A C, Greschuk L T, Curi N, Silva S H G. The fundamental of the effects of water, organic matter, and iron forms on the pXRF information in soil analyses [J]. CATENA, 2022, 210: 105868

[64]

Babos D V, Tadini A M, De Morais C P, Barreto B B, Carvalho M A R, Bernardi A C C, Oliveira P P A, Pezzopane J R M, Milori D M B P, Martin-Neto L. Laser-induced breakdown spectroscopy (LIBS) as an analytical tool in precision agriculture: Evaluation of spatial variability of soil fertility in integrated agricultural production systems [J]. CATENA, 2024, 239: 107914

[65]

Zhou W, Xiao J Y, Li H R, Chen Q, Wang T, Wang Q, Yue T X. Soil organic matter content prediction using Vis-NIRS based on different wavelength optimization algorithms and inversion models[J]. Journal of Soils and Sediments, 2023, 23(6): 2506-2517

[66]

Gu X H, Wang Y C, Sun Q, Yang G J, Zhang C. Hyperspectral inversion of soil organic matter content in cultivated land based on wavelet transform[J]. Computers and Electronics in Agriculture, 2019, 167: 105053

[67]

Kweon G, Maxton C. Soil organic matter sensing with an on-the-go optical sensor [J]. Biosystems Engineering, 2013, 115(1): 66-81

[68]

Yuan J, Wang X, Yan C X, Chen S B, Wang S R, Zhang J Q, Xu Z Y, Ju X P, Ding N, Dong Y Z, Zhang W F. Wavelength selection for estimating soil organic matter contents through the radiative transfer model[J]. IEEE Access, 2020, 8: 176286-176293

[69]

Jiang Q H, Chen Y Y, Guo L, Fei T, Qi K. Estimating soil organic carbon of cropland soil at different levels of soil moisture using VIS-NIR spectroscopy[J]. Remote Sensing, 2016, 8(9): 755

[70]

Tian Y C, Zhang J J, Yao X, Cao W X, Zhu Y. Laboratory assessment of three quantitative methods for estimating the organic matter content of soils in China based on visible/near-infrared reflectance spectra[J]. Geoderma, 2013, 202: 161-170

[71]

Zhang Z P, Ding J L, Zhu C M, Wang J Z, Ma G L, Ge X Y, Li Z S, Han L J. Strategies for the efficient estimation of soil organic matter in salt-affected soils through Vis-NIR spectroscopy: Optimal band combination algorithm and spectral degradation [J]. Geoderma, 2021, 382: 114729

[72]

Lazaar A, Mouazen A M, Hammouti K EL, Fullen M, Pradhan B, Memon M S, Andich K, Monir A. The application of proximal visible and near-infrared spectroscopy to estimate soil organic matter on the Triffa Plain of Morocco[J]. International Soil and Water Conservation Research, 2020, 8(2): 195-204

[73]

Yang M H, Xu D Y, Chen S C, Li H Y, Shi Z. Evaluation of machine learning approaches to predict soil organic matter and pH using vis-NIR spectra[J]. Sensors, 2019, 19(2): 263

[74]

Ward K J, Foerster S, Chabrillat S. Estimating Soil Organic Carbon using multitemporal PRISMA imaging spectroscopy data[J]. Geoderma, 2024, 450: 117025

[75]

Asgari N, Ayoubi S, Demattê J A M, Dotto A C. Carbonates and organic matter in soils characterized by reflected energy from 350-25000 nm wavelength[J]. Journal of Mountain Science, 2020, 17(7): 1636-1651

[76]

Bai Y, Yang W, Wang Z Y, Cao Y Y, Li M Z. Improving the estimation accuracy of soil organic matter based on the fusion of near-infrared and Raman spectroscopy using the outer-product analysis [J]. Computers and Electronics in Agriculture, 2024, 219: 108760

[77]

Cao Y Y, Yang W, Li H, Zhang H, Li M Z. Development of a vehicle-mounted soil organic matter detection system based on near-infrared spectroscopy and image information fusion [J]. Measurement Science and Technology, 2024, 35(4): 045501

[78]

Yang H, Kuang B, Mouazen A M. Quantitative analysis of soil nitrogen and carbon at a farm scale using visible and near infrared spectroscopy coupled with wavelength reduction[J]. European Journal of Soil Science, 2012, 63(3): 410-420

[79]

Yu B, Yan C X, Yuan J, Ding N, Chen Z W. Prediction of soil properties based on characteristic wavelengths with optimal spectral resolution by using Vis-NIR spectroscopy[J]. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, 2023, 293: 122452

[80]

Xie S G, Ding F J, Chen S G, Wang X, Li Y H, Ma K. Prediction of soil organic matter content based on characteristic band selection method[J]. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, 2022, 273: 120949

[81]

胡晓艳, 崔旭, 韩小平, 张志勇, 秦刚, 宋海燕. 基于修正系数法的抗水分干扰土壤有机质近红外预测模型研究[J]. 光谱学与光谱分析, 2019, 39(4): 1059-1062

[82]

Hu X Y, Cui X, Han X P, Zhang Z Y, Qin G, Song H Y. Study on soil organic matter prediction model based on moisture correction algorithm and near infrared spectroscopy[J]. Spectroscopy and Spectral Analysis, 2019, 39(4): 1059-1062 (in Chinese)

[83]

李佳怿, 张翼腾, 周宝闯, 翁海勇, 周蓓蓓, 叶大鹏, 瞿芳芳. 光谱技术结合水分校正与样本增广的棉田土壤盐分精准反演[J]. 农业工程学报, 2024, 40(1): 171-181

[84]

Li J Y, Zhang Y T, Zhou B C, Weng H Y, Zhou B B, Ye D P, Qu F F. Inversing soil salinity in cotton fields using spectroscopy sample augmentation and moisture correction[J]. Transactions of the Chinese Society of Agricultural Engineering, 2024, 40(1): 171-181 (in Chinese)

[85]

陈睿华, 王怡婧, 张俊华, 丁启东, 贾科利, 许鑫杰. 利用正交信号校正消除水分影响的银川平原土壤盐分高光谱反演[J]. 农业工程学报, 2023, 39(19): 122-130

[86]

Chen R H, Wang Y J, Zhang J H, Ding Q D, Jia K L, Xu X J. Hyperspectral inversion of soil salinity after correcting moisture effect in Yinchuan Plain using orthogonal signals[J]. Transactions of the Chinese Society of Agricultural Engineering, 2023, 39(19): 122-130 (in Chinese)

[87]

Ji W J, Li S, Chen S C, Shi Z, Viscarra Rossel R A, Mouazen A M. Prediction of soil attributes using the Chinese soil spectral library and standardized spectra recorded at field conditions[J]. Soil and Tillage Research, 2016, 155: 492-500

[88]

Cao Y, Bao N S, Liu S J, Zhao W, Li S M. Reducing moisture effects on soil organic carbon content prediction in visible and near-infrared spectra with an external parameter othogonalization algorithm[J]. Canadian Journal of Soil Science, 2020, 100(3): 253-262

[89]

Roudier P, Hedley C B, Lobsey C R, Viscarra Rossel R A, Leroux C. Evaluation of two methods to eliminate the effect of water from soil vis–NIR spectra for predictions of organic carbon [J]. Geoderma, 2017, 296: 98-107

[90]

Ji W, Viscarra Rossel R A, Shi Z. Improved estimates of organic carbon using proximally sensed vis–NIR spectra corrected by piecewise direct standardization [J]. European Journal of Soil Science, 2015, 66(4): 670-678

[91]

杨斌. 基于高光谱的土壤有机质反演及水分影响剔除优化建模[D]. 淮南: 安徽理工大学, 2022

[92]

Yang B. Optimization modeling of soil organic matter inversion and moisture effect elimination based on hyperspectral data[D]. Huainan: Anhui University of Science & Technology, 2022 (in Chinese)

[93]

Zhao W D, Wu Z L, Yin Z D, Li D S. Reducing moisture effects on soil organic carbon content estimation in vis-NIR spectra with a deep learning algorithm[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2023, 16: 7733-7748

[94]

Bao Y L, Meng X T, Ruan W M, Liu H J, Wang M C, Mouazen A M. Leveraging moisture elimination and hybrid deep learning models for soil organic carbon mapping with multi-modal remote sensing data[J]. International Journal of Applied Earth Observation and Geoinformation, 2025, 139: 104513

[95]

Singh S, Kasana S S. Estimation of soil properties from the EU spectral library using long short-term memory networks[J]. Geoderma Regional, 2019, 18: e00233

[96]

Gallios I, Tziolas N. Synergistic use of low-cost nir scanner and geospatial covariates to enhance soil organic carbon predictions using dual input deep learning techniques[C]. In: IGARSS 2024-2024 IEEE International Geoscience and Remote Sensing Symposium. Athens: IEEE, 2024: 1610-1613

[97]

Hong Y S, Chen S C, Hu B F, Wang N, Xue J, Zhou Z Q, Yang Y Y, Chen Y Y, Peng J, Liu Y L, Mouazen A M, Shi Z. Spectral fusion modeling for soil organic carbon by a parallel input-convolutional neural network[J]. Geoderma, 2023, 437: 116584.

[98]

Yang W Q, Li F S, Lyu S B, Zhang Q L, Zhao Y C. Quantitative analysis of heavy metals in soil via hierarchical deep neural networks with X-ray fluorescence spectroscopy[J]. Journal of Analytical Atomic Spectrometry, 2023, 38(9): 1830-1840

[99]

Liu H, Zhu Q H, Xia X M, Li M W, Huang D Y. Multi-feature optimization study of soil total nitrogen content detection based on thermal cracking and artificial olfactory system[J]. Agriculture, 2022, 12(1): 37

[100]

Jafarbiglu H, Pourreza A. A comprehensive review of remote sensing platforms, sensors, and applications in nut crops [J]. Computers and Electronics in Agriculture, 2022, 197: 106844

[101]

Ustin S L, Middleton E M. Current and near-term earth-observing environmental satellites, their missions, characteristics, instruments, and applications [J]. Sensors, 2024, 24(11): 3488

[102]

Zhang Z X, Zhu L X. A review on unmanned aerial vehicle remote sensing: Platforms, sensors, data processing methods, and applications [J]. Drones, 2023, 7(6): 398

[103]

Meng X T, Bao Y L, Wang Y A, Zhang X L, Liu H J. An advanced soil organic carbon content prediction model via fused temporal-spatial-spectral (TSS) information based on machine learning and deep learning algorithms [J]. Remote Sensing of Environment, 2022, 280: 113166

[104]

Tziolas N, Tsakiridis N, Ogen Y, Kalopesa E, Ben-Dor E, Theocharis J, Zalidis G. An integrated methodology using open soil spectral libraries and Earth Observation data for soil organic carbon estimations in support of soil-related SDGs[J]. Remote Sensing of Environment, 2020, 244: 111793

[105]

Wang S, Guan K Y, Zhang C H, Lee D, Margenot A J, Ge Y F, Peng J, Zhou W, Zhou Q, Huang Y Z. Using soil library hyperspectral reflectance and machine learning to predict soil organic carbon: Assessing potential of airborne and spaceborne optical soil sensing[J]. Remote Sensing of Environment, 2022, 271: 112914

[106]

Wang X P, Zhang F, Kung H T, Johnson V C. New methods for improving the remote sensing estimation of soil organic matter content (SOMC) in the Ebinur Lake Wetland National Nature Reserve (ELWNNR) in northwest China [J]. Remote Sensing of Environment, 2018, 218: 104-118

[107]

Hassan-Esfahani L, Torres-Rua A, Jensen A, McKee M. Assessment of surface soil moisture using high-resolution multi-spectral imagery and artificial neural networks [J]. Remote Sensing, 2015, 7(3): 2627-2646

[108]

Meng X T, Bao Y L, Liu J G, Liu H J, Zhang X L, Zhang Y, Wang P, Tang H T, Kong F C. Regional soil organic carbon prediction model based on a discrete wavelet analysis of hyperspectral satellite data[J]. International Journal of Applied Earth Observation and Geoinformation, 2020, 89: 102111

[109]

Wang X P, Zhang F, Ding J L, Kung H T, Latif A, Johnson V C. Estimation of soil salt content (SSC) in the Ebinur Lake Wetland National Nature Reserve (ELWNNR), Northwest China, based on a Bootstrap-BP neural network model and optimal spectral indices [J]. Science of the Total Environment, 2018, 615: 918-930

[110]

Wei L F, Zhang Y X, Lu Q K, Yuan Z R, Li H B, Huang Q B. Estimating the spatial distribution of soil total arsenic in the suspected contaminated area using UAV-Borne hyperspectral imagery and deep learning[J]. Ecological Indicators, 2021, 133: 108384

[111]

Zhang Y X, Wei L F, Lu Q K, Zhong Y F, Yuan Z R, Wang Z X, Li Z Q, Yang Y J. Mapping soil available copper content in the mine tailings pond with combined simulated annealing deep neural network and UAV hyperspectral images[J]. Environmental Pollution, 2023, 320: 120962

[112]

Yin L B, Yan S C, Li M, Liu W Z, Zhang S, Xie X Y, Wang X X, Wang W T, Chang S H, Hou F J. Enhancing soil moisture estimation in alfalfa root-zone using UAV-based multimodal remote sensing and deep learning [J]. European Journal of Agronomy, 2024, 161: 127366

[113]

Wang S, Guan K Y, Zhang C H, Zhou Q, Wang S B, Wu X C, Jiang C Y, Peng B, Mei W Y, Li K Y, Li Z Y, Yang Y, Zhou W, Huang Y Z, Ma Z W. Cross-scale sensing of field-level crop residue cover: Integrating field photos, airborne hyperspectral imaging, and satellite data [J]. Remote Sensing of Environment, 2023, 285: 113366

[114]

Tziolas N, Tsakiridis N, Ben-Dor E, Theocharis J, Zalidis G. Employing a multi-input deep convolutional neural network to derive soil clay content from a synergy of multi-temporal optical and radar imagery data [J]. Remote Sensing, 2020, 12(9): 1389

[115]

Jiang R Z, Sui Y Y, Zhang X, Lin N, Zheng X M, Li B Z, Zhang L, Li X K, Yu H Y. Estimation of soil organic carbon by combining hyperspectral and radar remote sensing to reduce coupling effects of soil surface moisture and roughness[J]. Geoderma, 2024, 444: 116874

[116]

Wang Y X, Qu Z Y, Yang W, Chen X, Qiao T. Inversion of soil salinity in the irrigated region along the southern bank of the Yellow River using UAV multispectral remote sensing[J]. Agronomy, 2024, 14(3): 523

[117]

Hosseini F S, Razavi-Termeh S V, Sadeghi-Niaraki A, Choi S M, Jamshidi M. Spatial prediction of physical and chemical properties of soil using optical satellite imagery: A state-of-the-art hybridization of deep learning algorithm [J]. Frontiers in Environmental Science, 2023, 11: 1279712

[118]

Wang Y B, Zhang X, Huang C P, Qi W C, Wang J N, Yang X K, Ding S T, Tao S Y. Spatial-convolution spectral-transformer interactive network for large-scale fast refined land cover classification and mapping based on ZY1-02D satellite hyperspectral imagery [J]. Remote Sensing, 2023, 15(13): 3269

[119]

Kühnlein L, Keller S. Evaluation of transformers and convolutional neural networks for high-dimensional hyperspectral soil texture classification[C]. In: 2022 12th Workshop on Hyperspectral Imaging and Signal Processing: Evolution in Remote Sensing (WHISPERS). Rome: IEEE, 2022: 1-5

[120]

Wei Z S, Miao L G, Peng J, Zhao T J, Meng L K, Lu H, Peng Z Q, Cosh M H, Fang B, Lakshmi V, Shi J C. Bridging spatio-temporal discontinuities in global soil moisture mapping by coupling physics in deep learning[J]. Remote Sensing of Environment, 2024, 313: 114371

[121]

Huang S Z, Zhang X, Chen N C, Ma H L, Fu P, Dong J Z, Gu X H, Nam W H, Xu L, Rab G, Niyogi D. A novel fusion method for generating surface soil moisture data with high accuracy, high spatial resolution, and high spatio-temporal continuity [J]. Water Resources Research, 2022, 58(5): e2021WR030827

[122]

Wei Y T, Lin L P, Li J, Yuan Q Q. A soil moisture downscaling residual dense network considering spatiotemporal relationship[C]. In: IGARSS 2023 - 2023 IEEE International Geoscience and Remote Sensing Symposium. Pasadena: IEEE, 2023: 3170-3173

[123]

Zhao L Y, Tan K, Wang X, Ding J W, Liu Z X, Ma H L, Han B. Hyperspectral feature selection for SOM prediction using deep reinforcement learning and multiple subset evaluation strategies[J]. Remote Sensing, 2023, 15(1): 127

[124]

Cho Y, Sheridan A H, Sudduth K A, Veum K S. Comparison of field and laboratory VNIR spectroscopy for profile soil property estimation[J]. Transactions of the ASABE, 2017, 60(5): 1503-1510

[125]

Sun W C, Liu S, Zhang X, Li Y. Estimation of soil organic matter content using selected spectral subset of hyperspectral data[J]. Geoderma, 2022, 409: 115653

[126]

Conway L S, Sudduth K A, Kitchen N R, Anderson S H. Repeatability of commercially available visible and near infrared proximal soil sensors[J]. Precision Agriculture, 2023, 24(3): 1014-1029

[127]

Bönecke E, Meyer S, Vogel S, Schröter I, Gebbers R, Kling C, Kramer E, Lück K, Nagel A, Philipp G, Gerlach F, Palme S, Scheibe D, Zieger K, Rühlmann J. Guidelines for precise lime management based on high-resolution soil pH, texture and SOM maps generated from proximal soil sensing data [J]. Precision Agriculture, 2021, 22(2): 493-523

[128]

Wang D, Yang W, Meng C, Cao Y Y, Li M Z. Research on vehicle-mounted soil electrical conductivity and moisture content detection system based on current–voltage six-terminal method and spectroscopy [J]. Computers and Electronics in Agriculture, 2023, 205: 107640

[129]

Li J Y, Ren P Q, Bi Y P, Tolulope Opeyemi A, Cui J, Gao Y, Ye D P, Zhou B B, Weng H Y. Ground-based mobile imaging for detecting salt stress of cotton seedlings in the field[J]. Computers and Electronics in Agriculture, 2024, 227: 109550

[130]

Milella A, Reina G, Nielsen M. A multi-sensor robotic platform for ground mapping and estimation beyond the visible spectrum[J]. Precision Agriculture, 2019, 20(2): 423-444

[131]

Li H, Song Y M, Wang Z Y, Li M Z, Yang W. Development of an online prediction system for soil organic matter and soil moisture content based on multi-modal fusion[J]. Computers and Electronics in Agriculture, 2024, 227: 109514

[132]

Wu Z J, Cui N B, Zhang W J, Yang Y N, Gong D Z, Liu Q S, Zhao L, Xing L W, He Q Y, Zhu S D, Zheng S S, Wen S L, Zhu B. Estimation of soil moisture in drip-irrigated Citrus orchards using multi-modal UAV remote sensing [J]. Agricultural Water Management, 2024, 302: 108972

[133]

Rukhovich D I, Koroleva P V, Rukhovich D D, Rukhovich A D. Recognition of the bare soil using deep machine learning methods to create maps of arable soil degradation based on the analysis of multi-temporal remote sensing data[J]. Remote Sensing, 2022, 14(9): 2224

[134]

Lavallee J M, Soong J L, Cotrufo M F. Conceptualizing soil organic matter into particulate and mineral-associated forms to address global change in the 21st century[J]. Global Change Biology, 2020, 26(1): 261-273

[135]

Lehmann J, Kleber M. The contentious nature of soil organic matter[J]. Nature, 2015, 528(7580): 60-68

[136]

Lehmann J, Hansel C M, Kaiser C, Kleber M, Maher K, Manzoni S, Nunan N, Reichstein M, Schimel J P, Torn M S, Wieder W R, Kögel-Knabner I. Persistence of soil organic carbon caused by functional complexity[J]. Nature Geoscience, 2020, 13(8): 529-534

[137]

Kallenbach C M, Grandy A S, Frey S D, Diefendorf A F. Microbial physiology and necromass regulate agricultural soil carbon accumulation[J]. Soil Biology and Biochemistry, 2015, 91: 279-290

[138]

Liang C, Amelung W, Lehmann J, Kästner M. Quantitative assessment of microbial necromass contribution to soil organic matter[J]. Global Change Biology, 2019, 25(11): 3578-3590

[139]

Lugato E, Lavallee J M, Haddix M L, Panagos P, Cotrufo M F. Different climate sensitivity of particulate and mineral-associated soil organic matter[J]. Nature Geoscience, 2021, 14: 295-300

基金资助

国家重点研发计划(2024YFD1500800)

中国农业大学2115人才工程联合资助

AI Summary AI Mindmap
PDF (2869KB)

550

访问

0

被引

详细

导航
相关文章

AI思维导图

/

〈 〉