Journal of Integrative Agriculture ›› 2026, Vol. 25 ›› Issue (6): 2607-2622.DOI: 10.1016/j.jia.2025.08.016

• • 上一篇    下一篇

基于盐渍化主导因素的分类建模策略提高了遥感反演精度

  

  • 收稿日期:2025-04-01 修回日期:2025-08-21 接受日期:2025-07-02 出版日期:2026-06-20 发布日期:2026-05-07

A classification modeling strategy based on dominant factors of salinization to enhance remote sensing inversion accuracy

Mengchao Zheng1, 2, Jianjun Zhang2#, Weini Wang4, Zhigang Qiao6, Junmei Liu4, Min Gong1, Xiaobin Li1, 3#, Hongyuan Zhang1, 3, Yuyi Li1, 3, Ningning Li5, Lin Yang2, Wenjuan Li1   

  1. 1 State Key Laboratory of Efficient Utilization of Arid and Semi-arid Arable Land in Northern China, Institute of Agricultural Resources and Regional Planning, Chinese Academy of Agricultural Sciences, Beijing 100081, China

    2 School of Land Science and Technology, China University of Geosciences (Beijing), Beijing 100083, China

    3 National Center of Technology Innovation for Comprehensive Utilization of Saline-alkali Land, Dongying 257000, China

    4 Ordos Agriculture and Animal Husbandry Ecology and Resource Protection Center, Ordos 017001, China

    5 College of Geography and Remote Sensing Sciences, Xinjiang University, Urumqi 830046, China

    6 Inner Mongolia Autonomous Region Agriculture and Animal Husbandry Technology Extension Center, Hohhot 010010, China

  • Received:2025-04-01 Revised:2025-08-21 Accepted:2025-07-02 Online:2026-06-20 Published:2026-05-07
  • About author:Mengchao Zheng, E-mail: zhengmengchao1999@163.com; #Correspondence Xiaobin Li, E-mail: lixiaobin01@caas.cn; Jianjun Zhang, E-mail: j.zhang@cugb.edu.cn
  • Supported by:

    This work was supported by the National Key Research and Development Program Young Scientists Project, China (2023YFD1901900), the Key Research and Development Program of China (NK2022180406-01), the Key Science and Technology Tasks of the Chinese Academy of Agricultural Sciences (CAAS-ZDRW202407-02), the Ordos Saline-alkali Land Comprehensive Utilization of Agriculture and Animal Husbandry Science and Technology Innovation Test Demonstration Project, China (2023-KJCX-01), the Ordos City Bureau of Agriculture and Animal Husbandry Major Scientific and Technological Innovation Project, China (2023ZDKJCX), and the Taishan Scholars Program, China (TSQN202312322).

PDF

摘要：

土壤盐渍化是土地退化的主要表现之一，严重威胁农业可持续发展。基于遥感的方法是目前盐渍化监测的首选。然而环境因素的空间异质性，严重限制了遥感建模过程对土壤盐分含量（SSC）-建模因子关系的准确捕捉，最终影响监测精度。本研究提出了一种基于盐渍化发生主导因素的分类建模思路，通过划分土壤质地和地表排水条件，分类构建了盐渍化遥感反演模型。结果表明：分类建模显著增强了建模过程对SSC-建模因子的关系的捕捉。相同的建模指标与建模方法在不同分类情景下的适宜性具有显著差异。三种建模方法中，随机森林算法（RF）整体结果更稳健。三种变量优选方法中，LightGBM 方法与三种机器学习方法耦合效果整体最好。分类建模策略显著提升了反演模型精度，与未分类建模（R2=0.62）相比，测试集（R2=0.77）R2最高提升24%；地表排水条件较差情景的独立建模效果最好，其最佳耦合模型训练集R2=0.82，测试集R2=0.77。这项研究能够为精准农业背景下的土壤盐渍化遥感监测提供有价值的参考。

Abstract:

Soil salinization represents a primary manifestation of land degradation and presents a significant threat to sustainable agricultural development.  Remote sensing-based methodologies currently constitute the preferred approach for salinization monitoring.  Environmental factors’ spatial heterogeneity substantially constrains the modeling process in accurately capturing the soil salt content (SSC)-modeling factor relationship, thereby affecting monitoring accuracy.  This study proposes a classification modeling framework based on dominant salinization factors, establishing distinct remote sensing inversion models through categorization of soil texture and surface drainage conditions.  Results indicate that classification modeling substantially improves the capture of SSC-modeling factor relationships.  The efficacy of identical modeling indicators and methods varies significantly across different classification scenarios.  Among the three modeling approaches, random forest demonstrates superior overall robustness.  Of the three variable selection methods, light gradient boosting machine (LightGBM) shows the strongest compatibility with the modeling approaches.  The classification strategy significantly enhances model accuracy: compared to non-classified modeling (R2V=0.62), the testing set R² increases by up to 24% (R2V=0.77).  Models under poor surface drainage category demonstrate optimal performance, with coupled models achieving R2C=0.82 (training set) and R2V=0.77 (testing set).  This research provides valuable insights for remote sensing monitoring of soil salinization in precision agriculture contexts.

Key words: soil salinization , remote sensing ,  spatial heterogeneity ,  machine learning ,  food security