Urban vegetation extraction is very important for urban biodiversity assessment and protection. However, due to the diversity of vegetation types and vertical structure, it is still challenging to extract vertical information of urban vegetation accurately with single remotely sensed data. Airborne light detection and ranging (LiDAR) can provide elevation information with high-precision, whereas hyperspectral data can provide abundant spectral information on ground objects. The complementary advantages of LiDAR and hyperspectral data could extract urban vegetation much more accurately. Therefore, a three-dimensional (3D) vegetation extraction workflow is proposed to extract urban grasses and trees at individual tree level in urban areas using airborne LiDAR and hyperspectral data. The specific steps are as follows: (1) airborne hyperspectral and LiDAR data were processed to extract spectral and elevation parameters, (2) random forest classification method and object-based classification method were used to extract the two-dimensional distribution map of urban vegetation, (3) individual tree segmentation was conducted on a canopy height model (CHM) and point cloud data separately to obtain three-dimensional characteristics of urban trees, and (4) the spatial distribution of urban vegetation and the individual tree delineation were assessed by validation samples and manual delineation results. The results showed that (1) both the random forest classification method and object-based classification method could extract urban vegetation accurately, with accuracies above 99%; (2) the watershed segmentation method based on the CHM could extract individual trees correctly, except for the small trees and the large tree groups; and (3) the individual tree segmentation based on point cloud data could delineate individual trees in three-dimensional space, which is much better than CHM segmentation as it can preserve the understory trees. All the results suggest that two- and three-dimensional urban vegetation extraction could play a significant role in spatial layout optimization and scientific management of urban vegetation.
Soil contamination by heavy metals is of great concern due to the increasing intensity of anthropogenic activities and requires a reliable approach to rapidly diagnose heavy metal pollution in soils. In this study conducted in northern Shandong Province, China, a total of 110 soil samples were collected to determine the zinc concentration and their visible and near-infrared reflectance spectroscopy was applied to diagnose zinc contaminated soils. Pearson’s correlation analysis and competitive adaptive reweighted sampling algorithm were compared to select active bands as independent variables. Partial least squares regression and particle swarm optimization support vector machine method were applied to estimate zinc concentration and a leave-one-out cross-validation was used to validate data sets. Results show that the obvious accumulations of zinc were in topsoil at the study area. Partial least squares regression estimates of zinc concentrations were improved with competitive adaptive reweighted sampling algorithm. The optimized support vector machine model, based on the first derivative of spectra, achieved the best accuracy for zinc concentration estimation with ratio of performance to deviation of 2.74. The method of active band extracting by competitive adaptive reweighted sampling algorithm was responsible for the improved model performance because competitive adaptive reweighted sampling algorithms determined active bands which can be useful for interpreting complex relations with soil compounds, such as soil organic matter, iron oxide, and clay minerals. In addition, the climate in specific research areas plays an important role in active band selection as a result of soil constituent changes. This study indicated that a combination model of competitive adaptive reweighted sampling algorithm and optimized support vector machine (R2 = 0.88; RMSE = 6.74, RPD = 2.74) can be recommended for estimating zinc concentrations in polluted soils and can provide improved understanding of the estimation mechanism and an alternative approach for diagnosing soil contamination.
Source apportionment of potentially toxic elements in soils is a critical step for devising soil sustainable management strategies. However, misjudgment or imprecision can occur when traditional statistical methods are applied to identify and apportion the sources. The main objective of the study was to develop a robust approach composed of the absolute principal component score/multiple linear regression (APCS/MLR) receptor model, positive matrix factorization (PMF) receptor model and geostatistics to identify and apportion sources of soil potentially toxic elements in typical industrial and mining city, eastern China. APCS/MLR and PMF were applied to provide robust factors with contribution rates. The geostatistics coupled with the variography and kriging methods was used to present factors derived from these two receptor models. The results indicated that mean concentrations of As, Cd, Cr, Cu, Hg, Ni, Pb and Zn exceeded the local background levels. Based on multivariate receptor models and geostatistics, we determined four sources of eight potentially toxic elements including natural source (parent material), agricultural practices, pollutant emissions (industrial, mining and traffic) and the atmospheric deposition of coal combustion, which accounted for 68%, 12%, 12% and 9% of the observed potentially toxic element concentrations, respectively. This study provides a reliable and robust approach for potentially toxic elements source apportionment in this particular industrial and mining city with a clear potential for future application in other regions.