Open AccessPredicting Lead and Nickel Contamination in Soil using Spectroradiometer
Author(s) -
B.S. Pawar,
Ratnadeep R. Deshmukh
Publication year - 2021
Publication title -
international journal of recent technology and engineering
Language(s) - English
Resource type - Journals
ISSN - 2277-3878
DOI - 10.35940/ijrte.a5758.0510121
Subject(s) - environmental science , vnir , soil test , spectroradiometer , soil water , partial least squares regression , nickel , contamination , soil contamination , soil science , environmental chemistry , remote sensing , reflectivity , hyperspectral imaging , chemistry , geology , mathematics , materials science , metallurgy , statistics , ecology , physics , biology , optics
In the geosciences, visible–near–short-wave infrared reflectance spectroscopy seems to have the capability to become a helpful technique for soil classification, mapping, and remote confirmation of soil characteristics and mineral composition. Focus on improving the spatial resolution of soil maps in order to better deal with localized problems like soil pollution. A variety of physio-chemical properties were measured in long-term spiked soils with a range of lead and nickel concentrations and also their spectral reflectance between 400 and 2500 nm at three different locations in the agricultural region of MIDC, Aurangabad, Maharashtra, India. Principle component analysis (PCA) used for feature extraction of soil were partial least squares regression (PLSR) method is used for classification. To measured amount of lead and nickel in soil sample, thirteen features of soil samples are calculated. The main aim of this study was to use statistical methods to calculate the lead and nickel concentrations in soil, as well as to assess the efficiency of VNIR-SWIR reflectance spectroscopy for heavy metal estimation in soil using the ASD FieldSpec4 Spectroradiometer. R2 = 0.96 provides the best precision for lead content and R2 = 0.95 for nickel content in soil, according to the findings. Lead and nickel have RMSEs of 3.396 and 2.680, respectively. The outcomes show that the proposed method is capable of accurately forecasting lead and nickel concentrations.