Heavy Metal Concentration Estimation for Different Farmland Soils Based on Projection Pursuit and LightGBM with Hyperspectral Images

Abstract

Heavy metal pollution in farmland soil threatens soil environmental quality. It is an important task to quickly grasp the status of heavy metal pollution in farmland soil in a region. Hyperspectral remote sensing technology has been widely used in soil heavy metal concentration monitoring. How to improve the accuracy and reliability of its estimation model is a hot topic. This study analyzed 440 soil samples from Sihe Town and the surrounding agricultural areas in Yushu City, Jilin Province. Considering the differences between different types of soils, a local regression model of heavy metal concentrations (As and Cu) was established based on projection pursuit (PP) and light gradient boosting machine (LightGBM) algorithms. Based on the estimations, a spatial distribution map of soil heavy metals in the region was drawn. The findings of this study showed that considering the differences between different soils to construct a local regression estimation model of soil heavy metal concentration improved the estimation accuracy. Specifically, the relative percent difference (RPD) of As and Cu element estimations in black soil increased the most, by 0.30 and 0.26, respectively. The regional spatial distribution map of heavy metal concentration derived from local regression showed high spatial variability. The number of characteristic bands screened by the PP method accounted for 10-13% of the total spectral bands, effectively reducing the model complexity. Compared with the traditional machine model, the LightGBM model showed better estimation ability, and the highest determination coefficients (R²) of different soil validation sets reached 0.73 (As) and 0.75 (Cu), respectively. In this study, the constructed PP-LightGBM estimation model takes into account the differences in soil types, which effectively improves the accuracy and reliability of hyperspectral image estimation of soil heavy metal concentration and provides a reference for drawing large-scale spatial distributions of heavy metals from hyperspectral images and mastering soil environmental quality.

Keywords: LightGBM; hyperspectral image; local regression estimation; projection pursuit; soil heavy metals.

Figure 1

Map of the study area. (a) Sihe Town and its surrounding farming area in Jilin Province, China; (b) sampling site distribution map in the study area; (c) soil type distribution map in the study area.

Figure 2

Soil sample pretreatment. (a) Sample plot diagram of soil sample collection; (b) soil sample collection diagram; (c) heavy metal concentration determination chart for the collected samples.

Figure 3

Random forest-supervised classification extraction of bare soil image element results.

Figure 4

Flow chart of hyperspectral image-based heavy metal concentration estimation in various soil types.

Figure 5

Spectral curves of different soils and correlation with heavy metal elements. (a) Black soil spectral curve; (b) Albic black soil spectral curve; (c) Albic soil spectral curve; (d) Meadow soil spectral curve.

Figure 6

The cumulative contribution rate of the feature information’s major components. Yellow dotted line is Black soil; green dotted line is Albic black soil; red dotted line is Albic soil; blue dotted line is Meadow soil; purple dotted line is All soil.

Figure 7

A comparison of heavy metal concentration estimation accuracy based on different soil types.

Figure 8

Mapping of soil heavy metal concentration for the study area. (a,b) Map of heavy metal distribution in different soil types; (c,d) map of heavy metal distribution in whole samples.

Figure 9

Measured and predicted values. (a) LightGBM of soil As concentration; (b) LightGBM of soil Cu concentration; (c) PP–LightGBM of soil As concentration; (d) PP–LightGBM of soil Cu concentration; (e) PP-ELM of soil As concentration; (f) PP-ELM of soil Cu concentration.

Figure 10

Estimation accuracy (RPD) of the PP–LightGBM soil heavy metal concentration model. (a) Estimation accuracy (RPD) of As by PP-LightGBM model; (b) Estimation accuracy (RPD) of Cu by PP-LightGBM model.

Figure 11

Spatial distribution of heavy metal concentration and topographic data overlay analysis. (a) As superposition analysis results; (b) Cu superposition analysis results; (c) slope calculation results of the study area; A–D represents the area where As elements gather; E–H represents the area where Cu elements gather.