No description available.

Thesis (Ph.D.)--University of Alabama at Birmingham, 2008. / Title from PDF title page (viewed on July 12, 2010). Includes bibliographical references.

No description available.

Soil visible and near-infrared spectroscopy provides a non-destructive, rapid and low-cost approach to quantify various soil physical and chemical properties based on their reflectance in the spectral range of 400–2500 nm. With an increasing number of large-scale soil spectral libraries established across the world and new space-borne hyperspectral sensors, there is a need to explore methods to extract informative features from reflectance spectra and produce accurate soil spectroscopic models using machine learning. Features generated from regional or large-scale soil spectral data play a key role in the quantitative spectroscopic model for soil properties. The Land Use/Land Cover Area Frame Survey (LUCAS) soil library was used to explore PLS-derived components and fractal features generated from soil spectra in this study. The gradient-boosting method performed well when coupled with extracted features on the estimation of several soil properties. Transfer learning based on convolutional neural networks (CNNs) was proposed to make the model developed from laboratory data transferable for airborne hyperspectral data. The soil clay map was successfully derived using HyMap imagery and the fine-tuned CNN model developed from LUCAS mineral soils, as deep learning has the potential to learn transferable features that generalise from the source domain to target domain. The external environmental factors like the presence of vegetation restrain the application of imaging spectroscopy. The reflectance data can be transformed into a vegetation suppressed domain with a force invariance approach, the performance of which was evaluated in an agricultural area using CASI airborne hyperspectral data. However, the relationship between vegetation and acquired spectra is complicated, and more efforts should put on removing the effects of external factors to make the model transferable from one sensor to another.:Abstract I Kurzfassung III Table of Contents V List of Figures IX List of Tables XIII List of Abbreviations XV 1 Introduction 1 1.1 Motivation 1 1.2 Soil spectra from different platforms 2 1.3 Soil property quantification using spectral data 4 1.4 Feature representation of soil spectra 5 1.5 Objectives 6 1.6 Thesis structure 7 2 Combining Partial Least Squares and the Gradient-Boosting Method for Soil Property Retrieval Using Visible Near-Infrared Shortwave Infrared Spectra 9 2.1 Abstract 10 2.2 Introduction 10 2.3 Materials and methods 13 2.3.1 The LUCAS soil spectral library 13 2.3.2 Partial least squares algorithm 15 2.3.3 Gradient-Boosted Decision Trees 15 2.3.4 Calculation of relative variable importance 16 2.3.5 Assessment 17 2.4 Results 17 2.4.1 Overview of the spectral measurement 17 2.4.2 Results of PLS regression for the estimation of soil properties 19 2.4.3 Results of PLS-GBDT for the estimation of soil properties 21 2.4.4 Relative important variables derived from PLS regression and the gradient-boosting method 24 2.5 Discussion 28 2.5.1 Dimension reduction for high-dimensional soil spectra 28 2.5.2 GBDT for quantitative soil spectroscopic modelling 29 2.6 Conclusions 30 3 Quantitative Retrieval of Organic Soil Properties from Visible Near-Infrared Shortwave Infrared Spectroscopy Using Fractal-Based Feature Extraction 31 3.1 Abstract 32 3.2 Introduction 32 3.3 Materials and Methods 35 3.3.1 The LUCAS topsoil dataset 35 3.3.2 Fractal feature extraction method 37 3.3.3 Gradient-boosting regression model 37 3.3.4 Evaluation 41 3.4 Results 42 3.4.1 Fractal features for soil spectroscopy 42 3.4.2 Effects of different step and window size on extracted fractal features 45 3.4.3 Modelling soil properties with fractal features 47 3.4.3 Comparison with PLS regression 49 3.5 Discussion 51 3.5.1 The importance of fractal dimension for soil spectra 51 3.5.2 Modelling soil properties with fractal features 52 3.6 Conclusions 53 4 Transfer Learning for Soil Spectroscopy Based on Convolutional Neural Networks and Its Application in Soil Clay Content Mapping Using Hyperspectral Imagery 55 4.1 Abstract 55 4.2 Introduction 56 4.3 Materials and Methods 59 4.3.1 Datasets 59 4.3.2 Methods 62 4.3.3 Assessment 67 4.4 Results and Discussion 67 4.4.1 Interpretation of mineral and organic soils from LUCAS dataset 67 4.4.2 1D-CNN and spectral index for LUCAS soil clay content estimation 69 4.4.3 Application of transfer learning for soil clay content mapping using the pre-trained 1D-CNN model 72 4.4.4 Comparison between spectral index and transfer learning 74 4.4.5 Large-scale soil spectral library for digital soil mapping at the local scale using hyperspectral imagery 75 4.5 Conclusions 75 5 A Case Study of Forced Invariance Approach for Soil Salinity Estimation in Vegetation-Covered Terrain Using Airborne Hyperspectral Imagery 77 5.1 Abstract 78 5.2 Introduction 78 5.3 Materials and Methods 81 5.3.1 Study area of Zhangye Oasis 81 5.3.2 Data description 82 5.3.3 Methods 83 5.3.3 Model performance assessment 85 5.4 Results and Discussion 86 5.4.1 The correlation between NDVI and soil salinity 86 5.4.2 Vegetation suppression performance using the Forced Invariance Approach 86 5.4.3 Estimation of soil properties using airborne hyperspectral data 88 5.5 Conclusions 90 6 Conclusions and Outlook 93 Bibliography 97 Acknowledgements 117

In der vorliegenden Arbeit werden die Möglichkeiten einer Nutzung thermal-infraroter Wellenlängen zur fernerkundlichen Erfassung/Quantifizierung von Bodenparametern vorgestellt. Die Studie basiert auf Bodenproben des Untersuchungsgebietes Mullewa, welches sich in einer semiariden Agrarregion im West-Australischen Weizengürtel befindet. Im Mittelpunkt der Arbeit steht die Bewertung des langwelligen Infrarots (LWIR), innerhalb des atmosphärischen Fensters zwischen 8 und 14 Mikrometer, bezüglich seines spektralen Potentials für die quantitative Ableitung des Ton- und Sandgehaltes sowie des Gehaltes an organischem Kohlenstoff (SOC). Zur Abschätzung der Effizienz wurden die Ergebnisse des LWIR einer Quantifizierung aus dem herkömmlich gebrauchten solar-reflektiven Wellenlängenbereichs zwischen 0,4 und 2,5 Mikrometer (VNIR-SWIR) gegenübergestellt. Mit verschiedenen Methoden der Laborspektroskopie wurden Bodenproben aus dem Untersuchungsgebiet im thermalen (Emissions-FTIR-Spektroskopie und direktional-hemisphärische Reflexions- (DHR) Spektroskopie)und im solar-reflektiven (Diffuse Reflexions-Spektroskopie) Wellenlängenbereich eingemessen und anschließend auf ihren Informationsgehalt hin untersucht. Die quantitative Modellierung der pedologischen Parameter aus den gemessenen spektralen Signaturen wurde mithilfe einer multivariaten Regressionsanalyse (Partial Least Squares Regression – PLSR) realisiert. Diese Grundlagenstudie konnte zeigen, dass die spektralen Voraussetzungen im LWIR für ein mögliches Monitoring der Bodenparameter mit thermalen Fernerkundungsdaten gegeben sind. Die Arbeit demonstriert darüber hinaus, dass für die Erfassung/Quantifizierung der Textur-Parameter (Sand- und Tongehalt) der relevante spektrale Informationsgehalt im LWIR deutlich höher ist als im VNIR-SWIR. / This study embraces the feasibility of using the thermal infrared wavelength region for future remote sensing applications to detect/quantify soil parameters. The research is based on soil samples from the semiarid agricultural area of Mullewa, located within the wheat belt of Western Australia. The main focus of this study is to assess the potential of the longwave infrared (LWIR), within the atmospheric window between 8 and 14 micrometer, to predict the content of sand, clay and organic carbon (SOC) in soils. The results are compared with predictions made with the traditionally used solar-reflective wavelength region (visible, VIS: 0.4 - 0.7 micrometer; near infrared, NIR: 0.7 - 1.1 micrometer; shortwave infrared, SWIR: 1.1 - 2.5 micrometer). Using laboratory spectroscopy, the Mullewa soil samples were measured, both in the thermal infrared (emission FTIR spectroscopy and directional hemispherical reflection (DHR) spectroscopy) and in the solar-reflective (diffuse reflection spectroscopy) wavelength region. This data was analyzed to determine the relevant content of information for the soil parameters. Multivariate regression analyses (partial least squares regression - PLSR) were used to quantitatively model the soil parameters from the spectral signatures. This basic research demonstrated that the spectral requirements in the LWIR are met for monitoring these soil parameters with thermal remote sensing instruments. Furthermore the study found that the relevant spectral information for the detection/quantification of the sand- and the clay content (textural parameters) is explicitly higher in the LWIR than in the VNIR-SWIR.

Submitted by Angela Maria de Oliveira (amolivei@uepg.br) on 2018-11-06T17:18:16Z No. of bitstreams: 2 license_rdf: 811 bytes, checksum: e39d27027a6cc9cb039ad269a5db8e34 (MD5) Giancarlo Rodrigues.pdf: 1835625 bytes, checksum: 84e769e19af35cc8103d542fe655e171 (MD5) / Made available in DSpace on 2018-11-06T17:18:16Z (GMT). No. of bitstreams: 2 license_rdf: 811 bytes, checksum: e39d27027a6cc9cb039ad269a5db8e34 (MD5) Giancarlo Rodrigues.pdf: 1835625 bytes, checksum: 84e769e19af35cc8103d542fe655e171 (MD5) Previous issue date: 2018-09-13 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / A estimativa do teor de nutrientes do solo por espectroscopia de refletância difusa é feita através de um modelo de predição, do qual seu desempenho determina a efetividade do método em realizá-la. Esse modelo é elaborado por técnicas que procuram correlacionar dados de refletância de uma coleção de amostras ao respectivo valor de referência obtido por análise química, ambos dispostos como atributos de um conjunto de dados. Não obstante, a quantidade de atributos desse conjunto é elevada – alta dimensionalidade – e nem todos são relevantes à predição do nutriente de interesse, logo elaborar um modelo a partir de um conjunto com essas características envolve uma série de complicações que prejudicam seu desempenho de predição. Uma das estratégias para contorná-las é manter no conjunto de dados apenas atributos relevantes à predição do nutriente de interesse, o que é feito através da Seleção de Subconjunto de Atributos (SSA), porém a maioria dos algoritmos que a executam não apresentam desempenho satisfatório ao manusear conjuntos de alta dimensionalidade. A literatura pertinente, por outro lado, constatou que o emprego de algoritmos evolutivos para SSA em conjuntos com essa característica fornece subconjuntos de qualidade num tempo de execução aceitável, logo o objetivo desta dissertação foi identificar com o apoio da metaheurística de Otimização por Enxame de Partículas – PSO – os comprimentos de onda da região do infravermelho visível e próximo relevantes à predição do teor de alumínio trocável de amostras de solo da região dos Campos Gerais. Para isso, a SSA foi configurada como um problema de otimização em que o objetivo foi minimizar o valor de AIC dos modelos elaborados pelo algoritmo de Regressão Linear Múltipla a partir dos subconjuntos candidatos. Ademais, sabendo da influência dos parâmetros do algoritmo no resultado final, primeiro foram investigados os valores ideais para número de iterações, tamanho do enxame e valor de limiar que proporcionaram a seleção dos melhores subconjuntos, depois estes foram validados num conjunto de dados independente e o melhor apontado. Nossos resultados sugerem que, para nosso cenário, 40 iterações, tamanho de enxame 20 e limiar 0,6 fornecem os melhores subconjuntos, porém o desempenho de predição do melhor modelo identificado ainda é passível de aprimoramento. A redução proporcionada pelo método adotado foi significativa e por conta disso essa abordagem é indicada para SSA em conjuntos de dados de espectroscopia. / The soil nutrient content estimation by diffuse reflectance spectroscopy is done through a prediction model whose performance determines the method effectiveness when performing it. This model is elaborated by techniques that try correlating a sample collection’s reflectance data to the respective reference value obtained through chemical analysis, both arranged as dataset attributes. Nevertheless, the dataset attributes amount is large – high dimensionality – and not all of them are relevant to the interest nutrient’s prediction, so elaborating a model from a dataset with these characteristics involves some complications that impact its prediction performance. A strategy to circumvent them is keeping only relevant attributes to the interest nutrient’s prediction, which is done through Feature Subset Selection (FSS), but the majority of algorithms that perform it do not operate satisfactorily when handling highdimensional sets. On the other hand, the pertinent literature found that employing evolutionary algorithms for FSS in high-dimensionality datasets provides quality subsets in an acceptable execution time, so this master thesis’ objective was to identify with Particle Swarm Optimization – PSO – metaheuristic support the relevant wavelengths of visible and near infrared region for exchangeable aluminum content prediction of Campos Gerais region soil samples. For this, the FSS was configured as an optimization problem which the objective was to minimize the AIC value of candidate subsets models elaborated by Multiple Linear Regression algorithm. In addition, knowing the algorithm parameters influence on its final result, first the ideal values for iterations number, swarm size and threshold value that provided the selection of best subsets were investigated, then these subsets were validated in an independent dataset and the best established. Our results suggest that in our scenario 40 iterations, swarm size 20 and threshold 0.6 provided the best subsets, but the prediction performance of the best model is amenable to improvement. The dimensionality reduction provided by the adopted method was significant, so this approach is recommended for FSS in spectroscopy datasets.

Soil spectroscopy has experienced a tremendous increase in soil property characterisation, and can be used not only in the laboratory but also from the space (imaging spectroscopy). Partial least squares (PLS) regression is one of the most common approaches for the calibration of soil properties using soil spectra. Besides functioning as a calibration method, PLS can also be used as a dimension reduction tool, which has scarcely been studied in soil spectroscopy. PLS components retained from high-dimensional spectral data can further be explored with the gradient-boosted decision tree (GBDT) method. Three soil sample categories were extracted from the Land Use/Land Cover Area Frame Survey (LUCAS) soil library according to the type of land cover (woodland, grassland, and cropland). First, PLS regression and GBDT were separately applied to build the spectroscopic models for soil organic carbon (OC), total nitrogen content (N), and clay for each soil category. Then, PLS-derived components were used as input variables for the GBDT model. The results demonstrate that the combined PLS-GBDT approach has better performance than PLS or GBDT alone. The relative important variables for soil property estimation revealed by the proposed method demonstrated that the PLS method is a useful dimension reduction tool for soil spectra to retain target-related information.

Soil spectroscopy has experienced a tremendous increase in soil property characterisation, and can be used not only in the laboratory but also from the space (imaging spectroscopy). Partial least squares (PLS) regression is one of the most common approaches for the calibration of soil properties using soil spectra. Besides functioning as a calibration method, PLS can also be used as a dimension reduction tool, which has scarcely been studied in soil spectroscopy. PLS components retained from high-dimensional spectral data can further be explored with the gradient-boosted decision tree (GBDT) method. Three soil sample categories were extracted from the Land Use/Land Cover Area Frame Survey (LUCAS) soil library according to the type of land cover (woodland, grassland, and cropland). First, PLS regression and GBDT were separately applied to build the spectroscopic models for soil organic carbon (OC), total nitrogen content (N), and clay for each soil category. Then, PLS-derived components were used as input variables for the GBDT model. The results demonstrate that the combined PLS-GBDT approach has better performance than PLS or GBDT alone. The relative important variables for soil property estimation revealed by the proposed method demonstrated that the PLS method is a useful dimension reduction tool for soil spectra to retain target-related information.

Visible and near-infrared diffuse reflectance spectroscopy has been demonstrated to be a fast and cheap tool for estimating a large number of chemical and physical soil properties, and effective features extracted from spectra are crucial to correlating with these properties. We adopt a novel methodology for feature extraction of soil spectroscopy based on fractal geometry. The spectrum can be divided into multiple segments with different step–window pairs. For each segmented spectral curve, the fractal dimension value was calculated using variation estimators with power indices 0.5, 1.0 and 2.0. Thus, the fractal feature can be generated by multiplying the fractal dimension value with spectral energy. To assess and compare the performance of new generated features, we took advantage of organic soil samples from the large-scale European Land Use/Land Cover Area Frame Survey (LUCAS). Gradient-boosting regression models built using XGBoost library with soil spectral library were developed to estimate N, pH and soil organic carbon (SOC) contents. Features generated by a variogram estimator performed better than two other estimators and the principal component analysis (PCA). The estimation results for SOC were coefficient of determination (R2) = 0.85, root mean square error (RMSE) = 56.7 g/kg, the ratio of percent deviation (RPD) = 2.59; for pH: R2 = 0.82, RMSE = 0.49 g/kg, RPD = 2.31; and for N: R2 = 0.77, RMSE = 3.01 g/kg, RPD = 2.09. Even better results could be achieved when fractal features were combined with PCA components. Fractal features generated by the proposed method can improve estimation accuracies of soil properties and simultaneously maintain the original spectral curve shape.

Soil spectroscopy is a promising technique for soil analysis, and has been successfully utilized in the laboratory. When it comes to space, the presence of vegetation significantly affects the performance of imaging spectroscopy or hyperspectral imaging on the retrieval of topsoil properties. The Forced Invariance Approach has been proven able to effectively suppress the vegetation contribution to the mixed image pixel. It takes advantage of scene statistics and requires no specific a priori knowledge of the referenced spectra. However, the approach is still mainly limited to lithological mapping. In this case study, the objective was to test the performance of the Forced Invariance Approach to improve the estimation accuracy of soil salinity for an agricultural area located in the semi-arid region of Northwest China using airborne hyperspectral data. The ground truth data was obtained from an eco-hydrological wireless sensing network. The relationship between Normalized Difference Vegetation Index (NDVI) and soil salinity is discussed. The results demonstrate that the Forced Invariance Approach is able to improve the retrieval accuracy of soil salinity at a depth of 10 cm, as indicated by a higher value for the coefficient of determination (R2). Consequently, the vegetation suppression method has the potential to improve quantitative estimation of soil properties with multivariate statistical methods.