Quantitative assessment of soil contaminants, particularly hydrocarbons, using reflectance spectroscopy

Abstract

Apparatus and method for efficiently assessing the results of reflectance spectroscopy on a soil sample to determine the presence of contaminants in the soil, by constructing a model based on analysis of known samples. The model may be constructed using an all possibilities approach and data mining techniques, on a range of samples, for example of different kinds of soil without pollutants and with different levels of pollutants. The Disclosure relates both to the construction of the model and to its use in the field in analyzing soil contaminants.

Description

RELATED APPLICATION[0001]This application claims the benefit of priority under 35 USC 119(e) of U.S. Provisional Patent Application No. 61 / 659,494 filed Jun. 14, 2012, the contents of which are incorporated herein by reference in their entirety.FIELD AND BACKGROUND OF THE INVENTION[0002]The present invention, in some embodiments thereof, relates to quantitative assessment of hydrocarbon contamination in soil using reflectance spectroscopy and more particularly but not exclusively to the quantitative assessment of hydrocarbon contamination using near infra-red spectral assessment and a modeling approach such as artificial neural networks, fuzzy logic, partial least squares, support vector machine, and metric learning.[0003]The term “hydrocarbon contamination” is intended to include all kinds of artificial organic pollutants in soil that can be identified by reflectance spectroscopy.[0004]Petroleum hydrocarbons are contaminants of great significance. The commonly used analytic method ...

AI Technical Summary

Benefits of technology

[0019]The present embodiments provide a method for efficiently assessing the results of reflectance spectroscopy on a soil sample to determine the presence of contaminants in the soil, by constructing a model based on analysis of known samples. The model may be constructed using an all possibilities approach and data mining techniques, on a range of samples, for example of different kinds of soil without pollutants and with different levels of pollutants. The present disclosure relates both to the construction of the model and to its use in the field in analyzing soil contaminants.

[0020]An embodiment is a way of efficiently assessing the results of reflectance spectroscopy on a soil sample to determine the presence of contaminants in the soil, by constructing a model based on analysis of known samples. An exemplary model is the partial least squares regression model referred to above, however, the model is constructed using an all possibilities pre-processing approach and data mining techniques as will be discussed hereinbelow, on samples of different kinds of soil without pollutants and with different levels of pollutants. It was noted by the present inventors that NIRA calibration processes can be strongly affected by the various preprocessing techniques, and thus a new “brute force” approach is suggested; in which all possibilities of common preprocessing methods are tested before the modeling stage is applied.

Problems solved by technology

Exposure to higher levels can cause extreme debilitation, loss of balance, and may even lead to coma, seizures and lethality.

Long term exposure is proven to cause changes in neurophysiological or psychological capacity and is further known to induce increased risk of lung, skin and bladder cancer alongside other carcinogenic effects (Hutcheson et al., 1996; Boffetta et al., 1997; Ritchie et al., 2001).

Not only was this method withdrawn by the EPA, but it is also problematic for various other reasons such as the need for skilled operators, the process length and cost, the difficulties in using it in situ, availability of the extracting solvent being very limited, the need for transporting samples to the laboratory etc.

The first study utilizing field collected samples, was not able to produce robust models but rather led to very low correlations (r=0.68) and large errors, probably due to the limited number of samples and problems with the analytic chemistry measurements done by the laboratory that produced inconsistent measurements (Malley et al., 1999).

Attempts at mapping hydrocarbons using the Landsat and Daedalus sensors in 1994 and 1995 failed, probably due to the limited spectral resolution of the sensors (multispectral sensors, 7 and 12 bands respectively) (Kühn and Hörig, 1995; Hörig et al., 2001).

This index is limited to very high signal to noise ratio sensors as well as other issues, such as problems with land cover, vegetation and high concentration detection levels (Kühn et al., 2004).

Due to the numerous combinations of preprocessing techniques as well as dataset divisions there is a lack of effective tools to allow reflectance spectroscopy methods to be used effectively in situ and so today it is not possible to provide an automated and optimized NRA modeling system for hydrocarbon contamination analysis in soils in a way which is rapid, accurate, and cost effective, solely from reflectance spectroscopy.

Images