文献类型： 外文期刊

作者： Huai, Shengchang ¹ ; Zhang, Qingyue ¹ ; Jin, Yuwen ¹ ; Yu, Weijia ³ ; Meersmans, Jeroen ² ; Wang, Shichao ⁴ ; Colinet, Gilles ² ; Lu, Changai ¹ ;

作者机构： 1.Chinese Acad Agr Sci, Inst Agr Resources & Reg Planning, Beijing, Peoples R China

2.Univ Liege, Gembloux Agrobio Tech, Passage Deportes 2, B-5030 Gembloux, Belgium

3.Shandong Acad Agr Sci, Inst Crop Germplasm Resources, Jinan, Shandong Provin, Peoples R China

4.Chinese Acad Sci, Inst Genet & Dev Biol, Shijiazhuang, Hebei, Peoples R China

关键词： Soil nutrients; Infrared spectroscopy; NPK estimation; Machine learning; Predictive models; Interference factors

期刊名称：TRAC-TRENDS IN ANALYTICAL CHEMISTRY （ 影响因子：12.0； 五年影响因子：13.2 ）

ISSN： 0165-9936

年卷期： 2025 年 185 卷

页码：

收录情况： SCI

摘要： Soil nitrogen (N), phosphorus (P), and potassium (K) are vital nutrients that underpin plant health and agricultural productivity. Precise and rapid monitoring of soil NPK levels is crucial for promoting sustainable farming practices. Infrared spectroscopy (IR) has emerged as a promising non-invasive technology for real-time soil nutrient assessment. However, current predictive models face challenges in achieving high precision and robustness due to the inherent heterogeneity of soil and variations in environmental conditions. This review critically examines the latest advancements in infrared spectroscopy for NPK estimation, emphasizing innovations in data acquisition, preprocessing, variable selection, and modelling techniques. The effects of soil heterogeneity and environmental factors on predictive accuracy are thoroughly evaluated, alongside advanced strategies proposed to address these limitations. Particular focus is placed on the role of soil components in influencing spectroscopy-based NPK estimation models, emphasizing the need for future research to refine characteristic NPK spectroscopy band selection. Furthermore, the development of correction sub-models that account for the interference patterns of soil components is recommended to improve model accuracy and stability. Infrared spectroscopy holds great promise for precision agriculture by enabling real-time soil nutrient management, thereby contributing to global food security.