A model optimization-based indirect inversion system and method for total phosphorus and total nitrogen in a water body

By integrating multi-source remote sensing data and optimizing machine learning models, key feature sets of total phosphorus and total nitrogen in water bodies are extracted, solving the problems of long cycle, high cost and limited coverage of traditional monitoring methods. This achieves efficient and accurate inversion of total phosphorus and total nitrogen, supporting water environment monitoring and management.

CN121580859BActive Publication Date: 2026-06-02NORTH CHINA INST OF AEROSPACE ENG
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
NORTH CHINA INST OF AEROSPACE ENG
Filing Date
2025-12-19
Publication Date
2026-06-02

AI Technical Summary

Technical Problem

Traditional monitoring of total phosphorus and total nitrogen in water bodies relies on manual sampling and laboratory analysis, which has problems such as long cycle, high cost and limited coverage. Moreover, the hyperparameters of existing models depend on experience and are prone to getting trapped in local optima. They ignore the intrinsic relationship between TP and TN and other water quality indicators. The limited data sources restrict the accuracy and performance of the models.

Method used

A model-optimized indirect inversion system for total phosphorus and total nitrogen in water bodies is adopted. By integrating multi-source remote sensing data from Sentinel-2 and Gaofen-5 01A satellites, exclusive inversion feature sets for key parameters such as chlorophyll a, transparency, and water temperature are extracted. The machine learning model is optimized by combining particle swarm optimization and adaptive moment estimation algorithms to achieve globally optimal parameter adjustment and construct a two-level parameter optimization mechanism of global search and fine-tuning.

Benefits of technology

It achieves comprehensive and accurate inversion of total phosphorus and total nitrogen, improves inversion accuracy and stability, and breaks through the dilemma of long cycle, high cost and limited coverage of traditional monitoring. It can reflect the spatial distribution characteristics of eutrophication in water bodies in a timely and comprehensive manner, and provide efficient and scientific technical support for water environment monitoring.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121580859B_ABST
    Figure CN121580859B_ABST
Patent Text Reader

Abstract

The application discloses a kind of water body total phosphorus total nitrogen indirect inversion system and method based on model optimization, it is related to water body remote sensing inversion field, by the remote sensing data of Sentinel-2 satellite and high five No.01A satellite, the image and measured water sample data of monitoring area are acquired, and pretreated;Extraction chlorophyll a, transparency and water temperature inversion feature set, respectively train to obtain each parameter inversion model using machine learning model, generate spatial distribution diagram and extract sampling point inversion value;Inversion value and measured total phosphorus total nitrogen data are merged to build original sample set, establish prediction model, optimize prediction model by particle swarm algorithm combined with adaptive moment estimation algorithm, form optimal prediction model;Chlorophyll a, transparency and water temperature characteristics are input to optimal model pixel by pixel, realize the inversion of the concentration of total phosphorus total nitrogen in monitoring area.The application effectively solves the problems of limited coverage, model easy to fall into local optimum and other problems of traditional method, significantly improves the inversion accuracy and stability.
Need to check novelty before this filing date? Find Prior Art