• JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
 
  Bookmark and Share
 
 
Master's Dissertation
DOI
https://doi.org/10.11606/D.11.2016.tde-09032016-120133
Document
Author
Full name
Arnaldo Barros e Souza
E-mail
Institute/School/College
Knowledge Area
Date of Defense
Published
Piracicaba, 2015
Supervisor
Committee
Dematte, Jose Alexandre Melo (President)
Ferreira, Tiago Osorio
Terra, Fabricio da Silva
Title in Portuguese
Diagnóstico hiperespectral da relação textural entre horizontes de solo: construindo índices
Keywords in Portuguese
Atributo diagnóstico
CP-AFD
Índice Espectral de Relação Textural (IERT)
MidIR ACP
Pedologia
PLSR
VisNIR-SWIR
Abstract in Portuguese
A adoção de medidas afinadas à conservação do solo não é apenas uma necessidade, há muito passou a ser estratégico desenvolver técnicas que preencham as lacunas de informação e maximizem o uso adequado do solo. A diferenciação textural entre horizontes de solos é um dos principais aspectos a ser considerado quando do planejamento do uso do solo, pois se relaciona estreitamente com fatores como o enraizamento, a percolação e disponibilidade de água e a susceptibilidade dos solos à erosão. Diante disso, objetivou-se desenvolver Índices Espectroscópicos de Relação Textural (IERT) que estabeleçam quantitativamente o grau de diferenciação textural em perfis de solos via espectroscopia VisNIR-SWIR (350-2.500 nm) e MidIR (4000 - 400 cm-1). Foram utilizados dados espectrais de 150 perfis de solos descritos morfologicamente pertencentes a nove municípios das regiões Sudeste e Centro-Oeste do Brasil. Três modelos espectrais, VisNIR-SWIR, MidIR e VisNIR-SWIR-MidIR foram associados aos teores analíticos de matéria orgânica dos horizontes A e B, totalizando nove modelos, afim de se verificar o potencial daqueles puramente espectrais. Todos os modelos foram estudados por análises de funções discriminantes (CP-AFD) e regressão por mínimos quadrados (PLSR). A construção dos IERT's, num total de 15, cinco para cada modelo espectral, considerou a diferença de reflectância em bandas específicas ou associações entre elas, as quais foram selecionadas com base no estudo qualitativo e quantitativo das curvas espectrais. De modo geral, a diferenciação entre horizontes é ditada pelas regiões espectrais associadas principalmente aos óxidos de Fe, minerais de argila e quartzo. É possível determinar o grau de diferenciação textural entre horizontes de perfis de solos com excelente acurácia através de dados espectrais com taxa de acerto global de até 100 %, R2 de 0,76 a 0,82 e RPD de 2,02 a 2,35 nos modelos puramente espectrais, com destaque à região MidIR. O uso de IERT's produz bons a excelentes resultados, com R2 de 0,71 a 0,80 e RDP de 1,84 a 2,21 para os melhores índices de cada região espectral. O uso de índices espectrais reduz a dependência de métodos estatísticos avançados e dá suporte ao desenvolvimento de equipamentos óticos que trabalhem em regiões espectrais específicas, reduzindo custos e maximizando a aplicabilidade da técnica. Estudos abrangentes e exaustivos são indispensáveis antes que novos métodos como este se estabeleçam, particularmente em ciência do solo, na qual o objeto de estudo é, por natureza, complexo e intrigante.
Title in English
Hyperspectral diagnostic of textural ratio between soil horizons: building indices
Keywords in English
Diagnostic attribute
MidIR
PC-DFA
PCA
Pedology
PLSR
Spectral Textural Ratio Indices (STRI)
VisNIR-SWIR
Abstract in English
The adoption of measures related to soil conservation is not only a necessity, long it has become strategic to develop techniques that meet the information deficits and maximize the appropriate use of land. The textural differentiation between soil horizons is one of the main aspects to be considered dealing with land use planning as it closely relates to factors such as rooting, percolation and water availability and the susceptibility of soils to erosion. Therefore, it was aimed to develop Spectral Textural Ratio Indices (STRI) that quantitatively establish the degree of textural differentiation in soil profiles through VisNIR-SWIR (350-2500 nm) and MidIR (4000- 400 cm-1) spectroscopy. Spectral data of 150 soil profiles morphologically described belonging to nine municipalities in the Southeast and Midwest of Brazil were used. Three spectral models, VisNIR-SWIR, MidIR and VisNIR-SWIR-MidIR were associated with analytical organic matter content of the A and B horizons, comprising nine models in order to verify the potential of those purely spectral. All models were studied by discriminant functions analysis (CP-DFA) and Partial Least Square Regression (PLSR). The construction of the STRIs, a total of 15, five for each spectral model, considered the reflectance difference in specific bands or associations between them, which were selected based on the qualitative and quantitative study of spectral curves. In general, the differentiation between horizons is dictated primarily by spectral regions associated with iron oxides, clay minerals and quartz. It is possible to determine the degree of textural differentiation between soil profile horizons with great accuracy through spectral data with overall accuracy rate of up to 100%, R2 from 0.76 to 0.82 and RPD from 2.02 to 2.35 in the purely spectral models, with emphasis on MidIR region. The use of STRIs produces good to excellent results, with R2 ranging from 0.71 to 0.80 and RPD ranging from 1.84 to 2.21 for the best rates of each spectral region. The use of spectral indices reduces reliance on advanced statistical methods and supports the development of optical devices that work in specific spectral regions, reducing costs and maximizing the applicability of the technique. Large and exhaustive studies are imperatives before new methods like this are established, particularly in soil science, in which the object of study is by nature complex and intriguing.
 
WARNING - Viewing this document is conditioned on your acceptance of the following terms of use:
This document is only for private use for research and teaching activities. Reproduction for commercial use is forbidden. This rights cover the whole data about this document as well as its contents. Any uses or copies of this document in whole or in part must include the author's name.
Publishing Date
2016-03-15
 
WARNING: Learn what derived works are clicking here.
All rights of the thesis/dissertation are from the authors
CeTI-SC/STI
Digital Library of Theses and Dissertations of USP. Copyright © 2001-2024. All rights reserved.