欢迎访问《中国农学通报》,

中国农学通报 ›› 2026, Vol. 42 ›› Issue (11): 195-201.doi: 10.11924/j.issn.1000-6850.casb2025-0552

• 农业信息/科技教育 • 上一篇    下一篇

基于无人机多光谱遥感的土壤有机碳可视化监测平台

王志坤1(), 李新举2, 胡晓1()   

  1. 1 山东农业大学信息科学与工程学院, 山东泰安 271018
    2 山东农业大学资源与环境学院, 山东泰安 271018
  • 收稿日期:2025-07-01 修回日期:2025-12-24 出版日期:2026-06-12 发布日期:2026-06-12
  • 通讯作者:
    胡晓,男,1984年出生,山东临沂人,副教授,博士研究生,主要从事农业定量遥感研究。E-mail:
  • 作者简介:

    王志坤,男,1999年出生,山东泰安人,在读硕士研究生,研究方向:土壤养分遥感监测与信息化研究。通信地址:271018 山东泰安泰山区岱宗大街61号 山东农业大学信息科学与工程学院,E-mail:

  • 基金资助:
    国家自然科学基金项目“高潜水位煤矿沉陷区土壤生态变化过程及碳循环机理研究项目”(42077446)

Visual Monitoring Platform of Soil Organic Carbon Based on Unmanned Aerial Vehicle Multispectral Remote Sensing

WANG Zhikun1(), LI Xinju2, HU Xiao1()   

  1. 1 College of Information Science and Engineering, Shandong Agricultural University, Taian, Shandong 271018
    2 College of Resources and Environment, Shandong Agricultural University, Taian, Shandong 271018
  • Received:2025-07-01 Revised:2025-12-24 Published:2026-06-12 Online:2026-06-12

摘要:

土壤有机碳(SOC)含量和分布信息对农业生产至关重要。本研究创新性地集成无人机多光谱遥感影像、Web开发技术、机器学习算法和数据库技术,构建了一套土壤有机碳可视化监测平台。该平台应用于山东省济宁市兖州区煤矿区复垦农田的监测,结果表明:(1)平台能够基于无人机多光谱遥感影像,利用机器学习算法快速、准确预测SOC含量。其中,轻量级梯度提升机(LightGBM)模型为最佳预测模型,建模集和验证集的决定系数(R2)分别为0.825和0.793,验证了平台预测的可靠性;(2)平台实现了SOC的数据统计、含量分级、地理分布等多维度可视化展示,方便用户直观了解土壤状况。研究结果可为田块尺度的农田养分精准管理提供高效、便捷的技术手段和科学依据。

关键词: 土壤有机碳, 无人机, 多光谱, Web开发, 机器学习, 监测平台

Abstract:

Obtaining information on the content and distribution of soil organic carbon (SOC) is of great significance for the development of agricultural production. In this study, we designed and developed a SOC visual monitoring platform by innovatively integrating unmanned aerial vehicle (UAV) multispectral remote sensing images, Web development technology, machine learning algorithms and database technology. The system was applied to monitor the reclaimed farmlands in mining area of Yanzhou District, Jining City, Shandong Province. The results show that: (1) SOC content can be predicted quickly and accurately using UAV multispectral remote sensing images and machine learning algorithms. The light gradient boosting machine (LightGBM) model is the best prediction model, with the coefficient of determination (R2) of the modeling set and the validation set being 0.825 and 0.793, respectively; and (2) the system realizes visualizations for SOC data statistics, content grading, geographic distribution, and more. Therefore, the results of the study can provide scientific reference for the nutrient management of farmland at the field scale.

Key words: soil organic carbon, unmanned aerial vehicle, multispectral, Web development, machine learning, monitoring platform

中图分类号: