Abstract:The use of hyperspectral remote sensing to monitor heavy metal pollution in crops has become an important part of remote sensing research. The difference in the amount of heavy metal content in the contaminated crop leaves mapped to the spectrum is weak, so it is challenging to dig sensitively the value information contained in it. In this paper, based on the spectrum of crop leaves, a pollution detection model of copper pollution vegetation index (CPVI) was proposed by combining multiple spectral feature bands to characterize the pollution degree of heavy metal copper on crops. Firstly, a pot experiment was conducted to add CuSO4 center dot 5H(2)O powder with different concentration gradients to the soil to simulate copper-contaminated soil environment and stress corn growth. The spectra of old, middle and new leaves at the ear of corn were collected, and the Cu2+ content and relative chlorophyll concentration in the leaves were determined. Then, using 58 randomly selected corn leaf spectra as experimental data, the spectral reflectances of the two groups of wavelengths lambda 1 and lambda 1 were selected in the wavelength range of 380-900 nm. The Pearson correlation coefficient between CPVI [lambda 1 , lambda 2 ] and Cu2+ content in the corresponding leaves was calculated, and the absolute value matrix of correlation characteristics was obtained. Secondly, according to the obtained correlation feature matrix, the characteristic band of 690 and 465 nm with high correlation coefficient was extracted and combined with the band 850 nm to establish the Copper pollution index of maize (CPVIm). After that, CPVIm index was verified based on 26 other groups of data, and Normalized difference vegetation index (NDVI) , MERIS terrestrial chlorophyll index (MTCI) and other conventional vegetation indexes were compared to verify the effectiveness and superiority of CPVIm . The results showed that the highest correlation coefficient between NDVI, MTCI, REP, DVI and Cu2+ content in leaves was 0. 68, and the lowest residual sum of squares was 70. 99. However, CPVIm was significantly negatively correlated with Cu2+ content in leaves. The correlation coefficient was -0. 80, and the residual sum of squares was 48. 52, which were better than conventional indexes such as NDVI and MTCI. It proved that CPVIm is more sensitive to heavy metal stress. At the same time, the robustness of CPVIm index was verified by using the spectral data of different varieties of maize in different years. The correlation coefficient of CPVIm and Cu2+ content were -0. 90 and -0. 96, respectively, which were significantly correlated. It shows that CPVIm is still suitable for detecting the pollution degree of different maize varieties. In addition, using Cu2+ content, CPVIm and chlorophyll relative concentration in maize leaves, a three-dimensional analysis model was constructed, which reflected the correlation between them from a spatial point of view. The CPVI detection model based on the combination of spectral characteristic bands can be used as a reference method to evaluate the pollution degree of heavy metals in crops. The CPVL, index based on this method can effectively identify the degree of heavy metal Cu2+ pollution in maize.

Prediction of Maize Copper Content in Copper-Poor Area Based on Machine Learning

Machine Learning Methods to Predict Cadmium (cd) Concentration in Rice Grain and Support Soil Management at a Regional Scale

Predicting Bioaccumulation of Potentially Toxic Element in Soil–Rice Systems Using Multi-Source Data and Machine Learning Methods: A Case Study of an Industrial City in Southeast China

Leveraging machine learning for sustainable cultivation of Zn-enriched crops in Cd-contaminated karst regions

Prediction of cadmium contents in rice grains from Quaternary sediment-distributed farmland using field investigations and machine learning

Predictive model for cadmium uptake by maize and rice grains on the basis of bioconcentration factor and the diffusive gradients in thin-films technique

Modelling Bioaccumulation of Heavy Metals in Soil-Crop Ecosystems and Identifying Its Controlling Factors Using Machine Learning.

Prediction Heavy Metals Accumulation Risk in Rice Using Machine Learning and Mapping Pollution Risk

Predicting soil available cadmium by machine learning based on soil properties

Simulation of the Availability Index of Soil Copper Content Using General Regression Neural Network

Combining Optical, Fluorescence, Thermal Satellite, and Environmental Data to Predict County-Level Maize Yield in China Using Machine Learning Approaches.

Machine learning-based prediction of cadmium pollution in topsoil and identification of critical driving factors in a mining area

Predicting China's Maize Yield Using Multi-Source Datasets and Machine Learning Algorithms

Improved mapping of heavy metals in agricultural soils using machine learning augmented with spatial regionalization indices

Hyperspectral Inversion of Heavy Metal Copper Content in Corn Leaves Based on DRS–XGBoost

Using machine learning to predict selenium and cadmium contents in rice grains from black shale-distributed farmland area

Application of cadmium prediction models for rice and maize in the safe utilization of farmland associated with tin mining in Hezhou, Guangxi, China

A Model on Detecting the Polluted Degree of Maize Leaves by Cu Pollution Vegetation Index

Predicting and Mapping the Mn Content in Soil Based on Cokriging

Machine learning method for predicting cadmium concentrations in rice near an active copper smelter based on chemical mass balance

Using Optimized Spectral Indices and Machine Learning Algorithms to Assess Soil Copper Concentration in Mining Areas