A Robust Fuzzy Neural Network Model for Soil Lead Estimation from Spectral Features

Rohollah Goodarzi,M Zoej,Mehdi Mokhtarzade

doi:10.3390/rs70708416

Rohollah Goodarzi, M Zoej + Show 1 more

Open Access

https://doi.org/10.3390/rs70708416

Copy DOI

Journal: Remote sensing	Publication Date: Jun 29, 2015
Citations: 17	License type: CC BY 4.0

Affiliation: K.N.Toosi University of Technology

Abstract

Soil lead content is an important parameter in environmental and industrial applications. Chemical analysis, the most commonly method for studying soil samples, are costly, however application of soil spectroscopy presents a more viable alternative. The first step in the method is usually to extract some appropriate spectral features and then regression models are applied to these extracted features. The aim of this paper was to design an accurate and robust regression technique to estimate soil lead contents from laboratory observed spectra. Three appropriate spectral features were selected according to information from other research as well as the spectrum interpretation of field collected soil samples containing lead. These features were then applied to common Multiple Linear Regression (MLR), Partial Least Square Regression (PLSR) and Neural Network (NN) regression models. Results showed that although NN had adequate accuracy, it produced unstable results (i.e., variation of response in different runs). This problem was addressed with application of a Fuzzy Neural Network (FNN) with a least square training strategy. In addition to the stabilized and unique response, the capability of the proposed FNN was proved in terms of regression accuracy where a Ratio of Performance to Deviation (RPD) of 8.76 was achieved for test samples.

Highlights

Accurate and reliable estimation of heavy metal contents in soil samples is important for environmental studies
The appropriate extraction of spectral features is reported followed by implementation in linear and non-linear regression models
MLP and Partial Least Square Regression (PLSR) were used for linear model evaluations and Artificial Neural Network (ANN) and Fuzzy Neural Network (FNN) were assessed for nonlinear regression comparison

Summary

Introduction

Accurate and reliable estimation of heavy metal contents in soil samples is important for environmental studies. Any increased concentration of such elements like lead (Pb) in soil presents a health threat to most creatures and humans [1,2,3]. Extraction of these elements is important for several industrial applications such as ammunition, burial vault liners, ceramic glazes, leaded class and crystal and water lines and pipes [4]. Pb contents in soils have been determined in a laboratory where prepared soil samples are subjected to expensive and time consuming analysis [2,5,6].

Objectives

Methods

Results

Conclusion