In soil environmental monitoring, the indirect spectral inversion of heavy metal pollution stands as a prominent focus of contemporary research. Heavy metals such as cadmium (Cd), lead (Pb), mercury (Hg), chromium (Cr), copper (Cu), zinc (Zn), and arsenic (As) do not inherently display distinct spectral absorption peaks in the visible to near-infrared spectrum. However, heavy metal ions undergo adsorption, complexation, or co-precipitation with soil organic matter, iron/manganese oxides, and clay minerals, altering the concentration and morphology of these spectrally active components. Consequently, this indirectly modifies the soil reflectance spectrum. By utilizing multispectral data to extract characteristic bands significantly correlated with heavy metal concentrations, and pairing these with algorithms such as Partial Least Squares Regression (PLSR), Support Vector Regression (SVR), or Random Forest (RF), researchers can achieve rapid estimation of heavy metal content and map its spatial distribution. Studies indicate that spectral bands associated with organic matter and iron oxides (such as those near 600–800 nm, 1400 nm, 1900 nm, and 2200 nm) exhibit high sensitivity to variations in heavy metal concentration. By extracting sensitive band reflectance and corresponding combination indices via multispectral UAV or satellite imagery, region-scale heavy metal predictive models can be established to assist in identifying pollution risk zones and tracking contamination sources. Compared to traditional grid-based sampling paired with laboratory chemical assays, multispectral remote sensing drastically lowers sampling density and testing costs, making it exceptionally suited for wide-area soil surveys around industrial enterprises, sewage irrigation zones, and downstream tailing storage facilities. The HG-MultiSP-800 lightweight, compact UAV multispectral camera, developed by Hagorun Technology Limited, integrates multiple agricultural and environmental application bands and adapts seamlessly to mainstream UAV platforms, supplying a flexible and highly efficient remote sensing data acquisition instrument for soil environmental monitoring. For the decoupling of multi-metal composite pollution, combining multispectral data with feature variable selection algorithms—such as the Successive Projections Algorithm (SPA) and Competitive Adaptive Reweighted Sampling (CARS)—enables the extraction of sensitive band combinations for individual metallic elements. This facilitates the construction of multi-component quantitative models for the synchronous estimation of diverse heavy metals in composite contaminated zones, yielding data support for source apportionment and risk mitigation.

Unlike heavy metals, specific organic contaminants—such as petroleum hydrocarbons and polycyclic aromatic hydrocarbons (PAHs)—contain functional groups like C-H and C=C, which generate direct spectral absorption features in the near-infrared band. In petroleum hydrocarbon pollutants, the C-H bonds exhibit characteristic absorption peaks near 1720 nm, 1760 nm, 2310 nm, and 2340 nm, providing explicit markers for multispectral technology to identify petroleum-contaminated soils. By extracting band reflectance values sensitive to organic pollution and constructing spectral indices (such as the hydrocarbon index), suspected contamination zones and relative pollution levels can be rapidly mapped under field conditions, guiding strategic on-site sampling grid design. In soil environmental investigations of oil field development zones, refinery peripheries, and gas station sites, drone-mounted multispectral cameras can rapidly generate high-resolution spatial distribution maps of contaminants. Compared to traditional grid sampling workflows characterized by long turnarounds and high costs, multispectral remote sensing screens several square kilometers for contamination within tens of minutes, pinpointing pipeline leakage hotspots, oily sludge dumping grounds, and oily wastewater infiltration pathways. The HG-MultiSP-800 multispectral camera from Hagorun Technology Limited features narrow-band filters and stable radiometric capture performance, allowing its acquired data products to be cross-validated with terrestrial spectral measurements to elevate the diagnostic accuracy of organic pollution mapping. It should be highlighted that the spectral detection of soil organic pollution is subject to interference from soil taxonomy, background soil organic matter, and moisture content. Fusing multispectral imagery with auxiliary data layers—such as topography and land-use data—and coupling this with targeted field verification sampling can effectively minimize false-positive rates and enhance pollution identification precision.

Soil salinization represents a major type of soil degradation affecting agricultural productivity and ecological stability. By extracting salinity-sensitive spectral indices (such as the Salinity Index [SI], Normalized Difference Salinity Index [NDSI], and Brightness Index [BI]) and integrating them with field-measured salt concentration data, regression models can be established to achieve rapid estimation of soil salt content and grade salinization severity. Saline ions (Na+, Ca2+, Mg2+, etc.) and their hydration processes alter the water molecule absorption intensity and peak shape in the near-infrared band (especially near the 1400 nm and 1900 nm water absorption bands). Concurrently, salt crystallization causes an overall elevation in soil surface reflectance. Multispectral imagery captures these fine spectral variations, making it highly applicable for dynamic monitoring in arid, semi-arid, and coastal saline soil regions. Beyond salinity, multispectral technology can be deployed for the rapid prediction of soil organic matter, cation exchange capacity, and moisture content. Soil organic matter exhibits a distinct spectral response in the 600–800 nm region, where its concentration correlates negatively with reflectance in the visible bands. Building organic matter predictive models from multispectral imagery allows for the acquisition of data across large batches of sample sites without relying on chemical analysis, significantly reducing laboratory analysis workloads. In the tracking and evaluation of contaminated site remediation efficacy, multi-temporal multispectral imagery monitors the spatio-temporal variations of soil physicochemical parameters during the cleanup process, helping judge the validity and required cycle of remediation measures. The HG-MultiSP-800 camera from Hagorun Technology Limited features a compact, lightweight design optimized for small-to-medium UAV integration. It operates alongside ground calibration panels for synchronous data capture to yield absolute reflectance products, delivering a streamlined, highly portable multispectral data acquisition solution for soil environmental monitoring workflows.

Interested in Advancing Multispectral Technology Applications?

Our engineering group delivers advanced technical consulting and integrated solutions

Hagorun Technology Limited  |  Focus · Dedication · Exploration · Vision