Integrating a Hybrid Back Propagation Neural Network and Particle Swarm Optimization for Estimating Soil Heavy Metal Contents Using Hyperspectral Data

Liu, Piao; Liu, Zhenhua; Hu, Yueming; Shi, Zhou; Pan, Yuchun; Wang, Lu; Wang, Guangxing

doi:10.3390/su11020419

Cited by 41 publications

(28 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For this purpose, several methods, such as correlation analysis, VIF, and random forest are available. Studies have also shown that the Boruta algorithm exhibits superior performance with a higher accuracy and smaller error rate compared to the conventional statistical methods [32,43,51]. However, in our experiment it was also found that there was collinearity among the spectral variables selected by the Boruta algorithm.…”

Section: Discussionsupporting

confidence: 43%

“…As a commercially developed area, Guangdong has abundant naturally occurring non-ferrous metals and rare metal resources, and has become one of the most heavily contaminated areas in China. In the area, 65 training soil samples (black points in Figure 1) and 15 validation soil samples (cyan points in Figure 1) were collected and located using the global positioning system (GPS) during 22-24 June 2015 [32]. The study area for testing the optimal hyperspectral estimation models was located in the Conghua district of Guangzhou city, Guangdong (Figure 2a).…”

Section: Study Area and Datamentioning

confidence: 99%

“…One of the five points was located at the plot center and other four points were allocated along the diagonal lines of the plot and with an equal distance between the points. The plots were located using GPS [32].…”

Section: Soil Samplesmentioning

confidence: 99%

See 2 more Smart Citations

Estimation of Soil Heavy Metal Content Using Hyperspectral Data

et al. 2019

Self Cite

View full text Add to dashboard Cite

Quickly and efficiently monitoring soil heavy metal content is crucial for protecting the natural environment and for human health. Estimating heavy metal content in soils using hyperspectral data is a cost-efficient method but challenging due to the effects of complex landscapes and soil properties. One of the challenges is how to make a lab-derived model based on soil samples applicable to mapping the contents of heavy metals in soil using air-borne or space-borne hyperspectral imagery at a regional scale. For this purpose, our study proposed a novel method using hyperspectral data from soil samples and the HuanJing-1A (HJ-1A) HyperSpectral Imager (HSI). In this method, estimation models were first developed using optimal relevant spectral variables from dry soil spectral reflectance (DSSR) data and field observations of soil heavy metal content. The relationship of the ratio of DSSR to moisture soil spectral reflectance (MSSR) with soil moisture content was then derived, which built up the linkage of DSSR with MSSR and provided the potential of applying the models developed in the laboratory to map soil heavy metal content at a regional scale using hyperspectral imagery. The optimal relevant spectral variables were obtained by combining the Boruta algorithm with a stepwise regression and variance inflation factor. This method was developed, validated, and applied to estimate the content of heavy metals in soil (As, Cd, and Hg) in Guangdong, China, and the Conghua district of Guangzhou city. The results showed that based on the validation datasets, the content of Cd could be reliably estimated and mapped by the proposed method, with relative root mean square error (RMSE) values of 17.41% for the point measurements of soil samples from Guangdong province and 17.10% for the Conghua district at the regional scale, while the content of heavy metals As and Hg in soil were relatively difficult to predict with the relative RMSE values of 32.27% and 28.72% at the soil sample level and 51.55% and 36.34% at the regional scale. Moreover, the relationship of the DSSR/MSSR ratio with soil moisture content varied greatly before the wavelength of 1029 nm and became stable after that, which linked DSSR with MSSR and provided the possibility of applying the DSSR-based models to map the soil heavy metal content at the regional scale using the HJ-1A images. In addition, it was found that overall there were only a few soil samples with the content of heavy metals exceeding the health standards in Guangdong province, while in Conghua the seriously polluted areas were mainly distributed in the cities and of 26 croplands. This study implies that the new approach provides the potential to map the content of heavy metals in soil, but the estimation model of Cd was more accurate than those of As and Hg.

show abstract

Section: Discussionsupporting

confidence: 43%

Section: Study Area and Datamentioning

confidence: 99%

See 1 more Smart Citation

Estimation of Soil Heavy Metal Content Using Hyperspectral Data

et al. 2019

Self Cite

View full text Add to dashboard Cite

show abstract

“…For example, Zheng et al [12] suggested that it is feasible to predict AS element contents in soils using reflectance spectral, and the model results of 4 nm + multiplicative scatter correction (MSC) + partial least squares regression (PLSR) were the best (R 2 = 0.711, residual predictive deviation (RPD) = 1.827); Cheng et al [13] reported that AS contents in surface soils were detectable using visible/near-infrared spectral, and Savitzky-Golay (SG)+PLSR had the best effect (R 2 = 0.75, RPD = 1.81). Many previous studies have investigated models for the estimation of AS content from visible and neat infrared (VNIR) hyperspectral [12][13][14][15][16][17][18][19][20]. PLSR was usually chosen as the estimation model [20].…”

Section: Introductionmentioning

confidence: 99%

Estimating Soil Arsenic Content with Visible and Near-Infrared Hyperspectral Reflectance

et al. 2020

View full text Add to dashboard Cite

Soil arsenic (AS) contamination has attracted a great deal of attention because of its detrimental effects on environments and humans. AS and inorganic AS compounds have been classified as a class of carcinogens by the World Health Organization. In order to select a high-precision method for predicting the soil AS content using hyperspectral techniques, we collected 90 soil samples from six different land use types to obtain the soil AS content by chemical analysis and hyperspectral data based on an indoor hyperspectral experiment. A partial least squares regression (PLSR), a support vector regression (SVR), and a back propagation neural network (BPNN) were used to establish a relationship between the hyperspectral and the soil AS content to predict the soil AS content. In addition, the feasibility and modeling accuracy of different interval spectral resampling, different spectral pretreatment methods, feature bands, and full-band were compared and discussed to explore the best inversion method for estimating soil AS content by hyperspectral. The results show that 10 nm + second derivative (SD) + BPNN is the optimum method to predict soil AS content estimation; R 2 v is 0.846 and residual predictive deviation (RPD) is 2.536. These results can expand the representativeness and practicability of the model to a certain extent and provide a scientific basis and technical reference for soil pollution monitoring. Sustainability 2020, 12, x FOR PEER REVIEW 3 of 19 polluting enterprises in the suburb, such as chemical plants, funeral homes, medical waste treatment 95 plants, and so on. These polluting enterprises are also direct discharge areas of wastewater, waste 96 gas, and waste slag in Weinan City. In recent years, heavy metal pollution caused by city expansion 97 has drastically affects the soil ecological environment and human health. 98 2.2. Acquisition and Processing of Soil Data 99 The results of previous field investigations indicated that the soil AS content in the mining area 100 is mainly affected by human activities and shows a large difference, and the suburban soil AS 101 content is mainly affected by land use types and shows a small difference. Therefore, from 102 November to December, 2018, 90 soil samples were collected in accordance with the uniform grid 103 method with a higher sampling density in the mining area (50 m) and a lower density in the suburb 104 (200 m). The 50 soil samples were collected from the mining area and the rest form the suburb. There105 were six members in the research group with unified training on the sampling method. First, sample 106 points ware selected on the satellite map, and then the six members were divided into two groups to 107 finish the sampling job in each area. A real-time kinematic (RTK) was used to precisely locate every 108 sampling location. Figure 1 shows the position of the sampling points. The sampling was conducted 109 at a depth of 0-20 cm. Stones, plant residues, and other large debris were removed from each fresh 110 sample, which were then mixed tho...

show abstract

“…There are also many methods for evaluating heavy metal pollution; traditional methods such as the pollution index method, the enrichment index method, the Nemerow index method, and the ecological risk index method have been widely used [24]. New intelligent methods such as neural networks have also been used in heavy metal pollution research [25]. Some scholars have proposed an modified eco-risk assessment method, and based on the cost-effective effects of relevant decision makers, proposed a risk-based comprehensive risk management policy [26,27].…”

Section: Introductionmentioning

confidence: 99%

Spatial distribution and ecological risk assessment of heavy metals in soil from the Raoyanghe Wetland, China

Wang

Sun

et al. 2019

PLoS ONE

View full text Add to dashboard Cite

Wetlands are recognized as one of the most important natural environments for humans. At the same time, heavy metal pollution has an important impact on wetlands. China's Raoyanghe Wetland is one of the most important natural wild species gene banks in China. Eight heavy metal elements (As, Cd, Cr, Cu, Hg, Ni, Pb, and Zn) in surface layer and deep layer soils were analyzed using statistical-, pollution index-, and Nemerow index-based methods, the Hakanson potential ecological risk index method, and principal component and cluster analyses. The results showed that the maximum concentrations of heavy metals exceeded the background values in the core area and buffer zone of the wetland, but the heavy metal content of the soils was generally low and did not exceed 30%. With the exception of Hg, heavy metal concentrations showed strong spatial differentiation. The differences between the surface layer and deep layer soils of the core area were smaller than in the buffer zone. With the exception of Cd, a clear vertical zonation in the buffer zone soils was observed, showing greater evidence of external influences in this zone than the core. With the exception of partial surface soils, which indicated a safe level of pollution in the core area, all other soils were classified as having a ‘mild’ level of pollution. Thus, the wetland is moderately polluted, with both the core area and the buffer zone presenting a low level of potential ecological risk. According to the results of the present study, heavy metal contaminants in the wetland soils were found to be derived mainly from the natural sources.

show abstract

Integrating a Hybrid Back Propagation Neural Network and Particle Swarm Optimization for Estimating Soil Heavy Metal Contents Using Hyperspectral Data

Cited by 41 publications

References 24 publications

Estimation of Soil Heavy Metal Content Using Hyperspectral Data

Estimation of Soil Heavy Metal Content Using Hyperspectral Data

Estimating Soil Arsenic Content with Visible and Near-Infrared Hyperspectral Reflectance

Spatial distribution and ecological risk assessment of heavy metals in soil from the Raoyanghe Wetland, China

Contact Info

Product

Resources

About