Not available / text

The improved spectral resolution of modern hyperspectral sensors provides a means for discriminating subtly different classes of on ground materials in remotely sensed images. However, in order to obtain statistically reliable classification results, the number of necessary training samples can increase exponentially as the number of spectral bands increases. Obtaining the necessary number of training signals for these high-dimensional datasets may not be feasible. The problem can be overcome by preprocessing the data to reduce the dimensionality and thus reduce the number of required training samples. In this thesis, three dimensionality reduction methods, all based on parametric projection pursuits, are investigated. These methods are the Sequential Parametric Projection Pursuits (SPPP), Parallel Parametric Projection Pursuits (PPPP), and Projection Pursuits Best Band Selection (PPBBS). The methods are applied to very high spectral resolution data to transform the hyperspectral data to a lower-dimension subspace. Feature extractors and classifiers are then applied to the lower-dimensional data to obtain target detection accuracies. The three projection pursuit methods are compared to each other, as well as to the case of using no dimensionality reduction preprocessing. When applied to hyperspectral data in a precision agriculture application, discriminating sicklepod and cocklebur weeds, the results showed that the SPPP method was optimum in terms of accuracy, resulting in a classification accuracy of >95% when using a nearest mean, maximum likelihood, or nearest neighbor classifier. The PPPP method encountered optimization problems when the hyperspectral dimensionality was very high, e.g. in the thousands. The PPBBS method resulted in high classification accuracies, >95%, when the maximum likelihood classifier was utilized; however, this method resulted in lower accuracies when the nearest mean or nearest neighbor classifiers were used. When using no projection pursuit preprocessing, the classification accuracies ranged between ~50% and 95%; however, for this case the accuracies greatly depended on the type of classifier being utilized.

The thesis deals with use of hyperspectral data from APEX and AISA sensors for detection and classification of anthropogenic materials in the areas of Čáslav, Rokytnice nad Jizerou and Harrachov. The main goal is to propose methodology for the detection and classification of roof materials and road surface materials based on established spectral libraries. Another goal is to evaluate applicability of spectral libraries for classification, to compare possibilities of hyperspectral data with larger and smaller spectral range and to create maps of anthropogenic materials above. The methodological approach including masks of anthropogenic materials for roads surface materials and roof materials creation, settings of four classifications algorithms (Linear Spectral Unmixing, Multiple endmember spectral mixture analysis, Spectral Angle Mapper, Spectral Information Divergence) parameters and assessment of classification results, is in the methodology part. The results are visualized and evaluated using overall accuracy and percentage of classified pixels. Finally the results are compared with existing studies and possible improvements for further research are proposed. Powered by TCPDF (www.tcpdf.org)

The accurate estimation of atmospheric water vapour and the subsequent derivation of surface spectral reflectance from hyperspectral VNIR-SWIR remotely sensed data is important for many applications. A number of algorithms have been developed for estimating water vapour content from remotely sensed hyperspectral data that do not require in-situ measurements. Two algorithms, the Continuum Interpolated Band Ratio (CIBR) and the Atmospheric Precorrected Differential Absorption (APDA) have proven to be highly effective at estimating atmospheric water vapour. Although highly successful, the two methods still exhibit unwanted or spurious results when challenging conditions are encountered. Such conditions include the estimation of atmospheric water vapour over dark targets, when uncorrected atmospheric aerosols are present and over surfaces with complex spectral signatures.A differential absorption method called the Transmittance Slope Ratio (TSR) has been developed that negates these problems. The TSR method is comprised of a weighted mean radiance that is defined between two atmospheric water absorption features which is divided by a reference channel radiance to produce a measurable ratio value. This, is turn, may be related to a reference curve, such that, the TSR value may be expressed as an atmospheric water vapour content. To test the TSR method over real terrains, AVIRIS and HyMap measured hyperspectral radiometric data were used. Three test sites were used in total with each site allowing different aspects of the water vapour estimation to be critically examined. The sites are, Jasper Ridge and Moffett Field in California and Brukunga in South Australia.The TSR method is found to significantly improve estimated atmospheric water vapour over dark targets (with less than 3.5 % error for reflectances as low as 0.5 %), improvement over nonlinear surfaces, and finally, ++ / improvement in water vapour estimation when atmospheric aerosol conditions are not well known. In the final case the TSR method is found to estimate atmospheric water vapour with an error of less than 2 % when a 5 km visibility is assumed to be 25 km. The final result is at least an order of magnitude better than the CIBR and APDA methods.

With the increase in atmospheric CO2 concentrations, and the resulting potential for climate change, there has been increasing research devoted to understanding the factors that determine the magnitude of CO2 fluxes and the feedback of ecosystem fluxes on climate. This thesis is an effort to investigate the feasibility of using alternate methods to measure and estimate the CO2 exchange rates in the northern mixed grass prairie. Specifically, the objectives are to evaluate the capability of using ground-level hyperspectral, and satellite-level multispectral data in the estimation of mid-season leaf CO2 exchange rates as measured with a chamber, in and around Grasslands National Park (GNP), Saskatchewan. Data for the first manuscript was collected during June of 2004 (the approximate period for peak greenness for the study area). Spectral reflectance and CO2 exchange measurements were collected from 13 sites in and around GNP. Linear regression showed that the Photochemical Reflectance Index (PRI) calculated from hyperspectral ground-level data explained 46% of the variance seen in the CO2 exchange rates. This indicates that the PRI, which has traditionally been used only in laboratory conditions to predict CO2 exchange, can also be applied at the canopy level in grassland field conditions. <p>The focus of the second manuscript is to establish if the relationship found between ground-level hyperspectral data and leaf CO2 exchange is applicable to satellite-level derived vegetation indices. During June of 2005, biophysical and CO2 exchange measurements were collected from 24 sites in and around GNP. A SPOT satellite image was obtained from June 22, midway through the field data collection. Cubic regression showed that Normalized Difference Vegetation Index (NDVI) explained 46% of the variance observed in the CO2 exchange rates. To our knowledge, this is the first time that a direct correlation between satellite images and leaf CO2 fluxes has been shown within the grassland biome.

With the increase in atmospheric CO2 concentrations, and the resulting potential for climate change, there has been increasing research devoted to understanding the factors that determine the magnitude of CO2 fluxes and the feedback of ecosystem fluxes on climate. This thesis is an effort to investigate the feasibility of using alternate methods to measure and estimate the CO2 exchange rates in the northern mixed grass prairie. Specifically, the objectives are to evaluate the capability of using ground-level hyperspectral, and satellite-level multispectral data in the estimation of mid-season leaf CO2 exchange rates as measured with a chamber, in and around Grasslands National Park (GNP), Saskatchewan. Data for the first manuscript was collected during June of 2004 (the approximate period for peak greenness for the study area). Spectral reflectance and CO2 exchange measurements were collected from 13 sites in and around GNP. Linear regression showed that the Photochemical Reflectance Index (PRI) calculated from hyperspectral ground-level data explained 46% of the variance seen in the CO2 exchange rates. This indicates that the PRI, which has traditionally been used only in laboratory conditions to predict CO2 exchange, can also be applied at the canopy level in grassland field conditions. <p>The focus of the second manuscript is to establish if the relationship found between ground-level hyperspectral data and leaf CO2 exchange is applicable to satellite-level derived vegetation indices. During June of 2005, biophysical and CO2 exchange measurements were collected from 24 sites in and around GNP. A SPOT satellite image was obtained from June 22, midway through the field data collection. Cubic regression showed that Normalized Difference Vegetation Index (NDVI) explained 46% of the variance observed in the CO2 exchange rates. To our knowledge, this is the first time that a direct correlation between satellite images and leaf CO2 fluxes has been shown within the grassland biome.

The thesis deals with lignin and soluble phenolic determination in Norway spruce foliace using hyperspectral data. A literature overview is focused on the analysis of lignin and soluble phenolics. The practical part focuses on the determination of wavelenghts intervals which are suitable for the detection of lignin and soluble phenolics. There is applied regression analysis for the determination of relationship between the foliage spectra and the content of biochemical substances for the chosen spektrum intervals. Indexes NDLI, mNDLI and RLI were than calculated. HyMap hyperspectral airborne images from 2009 and 2010 for the area of Sokolov, spectral curves of dry matter and fresh branches of Norway spruce and laboratory determination of lignin and soluble phenolics content were the inputs for the analyses. Maps of lignin content in Norway spruce are the final output of the work. Keywords: Norway spruce (Picea Abies), lignin, soluble phenolics, PLS (partial least square) method, multiple Stepwise regression, NDLI

In this dissertation, novel techniques for hyperspectral classification and signal reconstruction from random projections are presented. A classification paradigm designed to exploit the rich statistical structure of hyperspectral data is proposed. The proposed framework employs the local Fisher’s discriminant analysis to reduce the dimensionality of the data while preserving its multimodal structure, followed by a subsequent Gaussianmixture- model or support-vector-machine classifier. An extension of this framework in a kernel induced space is also studied. This classification approach employs a maximum likelihood classifier and dimensionality reduction based on a kernel local Fisher’s discriminant analysis. The technique imposes an additional constraint on the kernel mapping—it ensures that neighboring points in the input space stay close-by in the projected subspace. In a typical remote sensing flow, the sender needs to invoke an appropriate compression strategy for downlinking signals (e.g., imagery to a base station). Signal acquisition using random projections significantly decreases the sender-side computational cost, while preserving useful information. In this dissertation, a novel class-dependent hyperspectral image reconstruction strategy is also proposed. The proposed method employs statistics pertinent to each class as opposed to the average statistics estimated over the entire dataset, resulting in a more accurate reconstruction from random projections. An integrated spectral-spatial model for signal reconstruction from random projections is also developed. In this approach, spatially homogeneous segments are combined with spectral pixel-wise classification results in the projected subspace. An appropriate reconstruction strategy, such as compressive projection principal component analysis (CPPCA), is employed individually in each category based on this integrated map. The proposed method provides better reconstruction performance as compared to traditional methods and the class-dependent CPPCA approach.

<p>High-throughput phenotyping using high spatial, spectral, and temporal resolution remote sensing (RS) data has become a critical part of the plant breeding chain focused on reducing the time and cost of the selection process for the “best” genotypes with respect to the trait(s) of interest. In this study, the potential of accurate and reliable sorghum biomass prediction using hyperspectral and LiDAR data acquired by sensors mounted on UAV platforms is investigated. Experiments comprised multiple varieties of grain and forage sorghum, including some photoperiod sensitive varieties, providing an opportunity to evaluate a wide range of genotypes and phenotypes. </p><p>Feature extraction is investigated, where various novel features, as well as traditional features, are extracted directly from the hyperspectral imagery and LiDAR point cloud data and input to classical machine learning (ML) regression based models. Predictive models are developed for multiple experiments conducted during the 2017, 2018, and 2019 growing seasons at the Agronomy Center for Research and Education (ACRE) at Purdue University. The impact of the regression method, data source, timing of RS and field-based biomass reference data acquisition, and number of samples on the prediction results are investigated. R2 values for end-of-season biomass ranged from 0.64 to 0.89 for different experiments when features from all the data sources were included. Using geometric based features derived from the LiDAR point cloud and the chemistry-based features extracted from hyperspectral data provided the most accurate predictions. The analysis of variance (ANOVA) of the accuracies of the predictive models showed that both the data source and regression method are important factors for a reliable prediction; however, the data source was more important with 69% significance, versus 28% significance for the regression method. The characteristics of the experiments, including the number of samples and the type of sorghum genotypes in the experiment also impacted prediction accuracy. </p><p>Including the genomic information and weather data in the “multi-year” predictive models is also investigated for prediction of the end of season biomass. Models based on one and two years of data are used to predict the biomass yield for the future years. The results show the high potential of the models for biomass and biomass rank predictions. While models developed using one year of data are able to predict biomass rank, using two years of data resulted in more accurate models, especially when RS data, which encode the environmental variation, are included. Also, the possibility of developing predictive models using the RS data collected until mid-season, rather than the full season, is investigated. The results show that using the RS data until 60 days after sowing (DAS) in the models can predict the rank of biomass with R2 values of around 0.65-0.70. This not only reduces the time required for phenotyping by avoiding the manual sampling process, but also decreases the time and the cost of the RS data collections and the associated challenges of time-consuming processing and analysis of large data sets, and particularly for hyperspectral imaging data.</p><p>In addition to extracting features from the hyperspectral and LiDAR data and developing classical ML based predictive models, supervised and unsupervised feature learning based on fully connected, convolutional, and recurrent neural networks is also investigated. For hyperspectral data, supervised feature extraction provides more accurate predictions, while the features extracted from LiDAR data in an unsupervised training yield more accurate prediction. </p><p>Predictive models based on Recurrent Neural Networks (RNNs) are designed and implemented to accommodate high dimensional, multi-modal, multi-temporal data. RS data and weather data are incorporated in the RNN models. Results from multiple experiments focused on high throughput phenotyping of sorghum for biomass predictions are provided and evaluated. Using proposed RNNs for training on one experiment and predicting biomass for other experiments with different types of sorghum varieties illustrates the potential of the network for biomass prediction, and the challenges relative to small sample sizes, including weather and sensitivity to the associated ground reference information.</p>

To support the growing human population, plant phenotyping technologies must innovate to rapidly interpret hyperspectral (HS) data into genetic inferences for plant breeders and managers. While pigment and nutrient concentrations within canopies are known to be vertically non-uniform, these chemical distributions as sources of HS noise are not universally addressed in scaling leaf information to canopy data nor in detecting spectral plant health traits. <br>In this project, soybeans (Glycine Max, cultivar Williams 82) were imaged with a Spectra Vista Corporation (SVC) HR-1024 spectroradiometer (350-2500 nm) at the highest five node positions. The samples were subjected to nitrogen and drought stress in factorial design (n=12) that was validated via relative water content (RWC) and PLS Regression of photopigments (chlorophyll a, chlorophyll b, lutein, neoxanthin, violaxanthin, and zeaxanthin in mg/g DW) and N concentration (%) for each imaged tissue. Welch’s ANOVA and Tamhane’s T2 post-hoc testing quantified spectral heterogeneity with respect to treatments and node positions through spectral angle measurements (SAMs) and percent NDVI difference. Drought-stressed samples had the lowest SAM between node positions compared to other treatments, and SAM node comparisons were greatest when including the highest sampled tissues. Taking ratios of NDVI between node positions proved more statistically effective at discerning between all factorial treatments than individual leaf NDVI values. Finally, intra-canopy spectral heterogeneity was exploited by training Linear Discriminant Analysis (LDA) classifiers on relative reflectance between node positions, tuning for the F1-Score. A classifier built on Node 1 vs. Node 3 reflectance outperformed in class-specific accuracies compared to analogous models trained on point-view data. Accounting for intra-canopy spectral variability is an opportunity to develop more comprehensive phenotyping tools for plant breeders in a world with rapidly rising agricultural demand.<br><br>