Multi-model collaborative inversion method used for water quality parameters and based on deterministic set modeling

Abstract

The invention relates to the field of remote sensing monitoring of water quality, in particular to a multi-model collaborative inversion method used for water quality parameters and based on deterministic set modeling. According to the method, sampling points are distributed in a research area, and water quality parameter concentration and corresponding hyperspectral remote sensing reflectivity of the sampling points are acquired; actually measured hyperspectral remote sensing reflectivity of a water body is preprocessed, and multiple water quality parameter inversion models are constructed by selecting preprocessed hyperspectral remote sensing reflectivity with better correlation with the water quality parameter concentration or band combination; weight of each single model is determined with an entropy weight method and a set pair principle, weighted summation is performed on inversion results of the single models, and multi-model collaborative inversion of the water quality parameters is realized. With adoption of the method, multi-model collaborative inversion of the water quality parameters can be realized with the two deterministic set modeling methods including the entropy weight method and the set pair principle by integrating characteristics of different water quality parameter inversion models, and the stability of the water quality parameter inversion results is improved.

Description

technical field [0001] The invention relates to the field of water quality remote sensing monitoring, in particular to a multi-model cooperative inversion method of water quality parameters based on deterministic set modeling. Background technique [0002] With the increasingly serious problem of water pollution, water quality monitoring has become a major issue that must be solved for the sustainable development of society and economy, especially in inland water bodies, whose water quality has threatened the production and domestic water safety of the people. , Rapid monitoring is of great significance to ensure national water safety. Conventional water quality monitoring mostly uses laboratory analysis methods. Although the monitoring accuracy is high, it is time-consuming and laborious and can only obtain the water quality status on the monitoring section, which is difficult to meet the needs of large-scale, multi-temporal dynamic monitoring of water quality; remote sensi...

AI Technical Summary

Problems solved by technology

The empirical model is a water quality inversion model based on multispectral data sources. This model is relatively simple and has the most applications, but its accuracy is limited. With the development of hyperspectral technology, semi-empirical and semi-analytical models are increasingly used in water quality In remote sensing, due to the semi-empirical / semi-analytical model has a certain physical meaning, the accuracy of the model has been improved; although the analytical model has a clear physical meaning, due to too many input parameters, the model mechanism is not very clear, and the inherent optical parameters of the water body are difficult to obtain Due to the limitations of other factors, analysis models are rarely used in practice; machine learning models have strong nonlinear simulation capabilities, and show unique advantages in water quality remote sensing inversion in complex water environments, but the models are relatively complex and prone to overfitting , the generalization ability needs to be further improved

Due to the complexity of the inland water environment, the above-mentioned single inversion model of water quality parameters is often inconsistent in modeling and verification accuracy, and it is difficult to achieve high accuracy at the same time, and the model stability is poor

Images