Theses and Dissertations at Montana State University (MSU)
Permanent URI for this communityhttps://scholarworks.montana.edu/handle/1/732
Browse
6 results
Search Results
Item Spectral processing for algae monitoring and mapping (SPAMM): remote sensing methodologies for river ecology(Montana State University - Bozeman, College of Engineering, 2024) Logan, Riley Donovan; Chairperson, Graduate Committee: Joseph A. Shaw; This is a manuscript style paper that includes co-authored chapters.Inland water quality is a growing concern to public health, riparian ecosystems, and recreational uses of our waterways. Many modern water quality programs include measures of the presence and abundance of harmful and nuisance algae. In southwestern Montana, large blooms of the nuisance algae, Cladophora glomerata, have become common in the Upper Clark Fork River due to a combination of warming water temperatures, naturally high phosphorus levels, and an influx of contaminants through wastewater and anthropogenic activity along its banks. To improve understanding of bloom dynamics, such as algal biomass and percent algae cover, and their effects on water quality, a UAV-based hyperspectral imaging system was used to monitor several locations along the Upper Clark Fork River. Image data were collected across the spectral range of 400 - 1000 nm with 2.1 nm spectral resolution during field sampling campaigns across the entirety of the project, beginning in 2019 and ending in 2023. In this dissertation, methodologies for monitoring water quality were developed. These methods include estimating benthic algal pigment abundance using spectral band ratios achieving R 2 values of up to 0.62 for chlorophyll alpha and 0.96 for phycocyanin; creating spatial algae distribution maps and estimating percent algae cover using machine learning classification algorithms with accuracies greater than 99%; combining spatial algae distribution maps and improved pigment estimation using machine learning regression algorithms for creating chlorophyll alpha abundance maps, achieving an R 2 of 0.873, while also comparing abundance values to Montana water quality thresholds; and identifying salient wavelengths for monitoring and mapping algae to inform the design of a low-cost and compact multispectral imager. Throughout all field campaigns, significant spatial variations in algal growth within each river reach and frequent violations of current water quality standards were observed, demonstrating the need for high-spatial resolution monitoring techniques to be incorporated in current water quality monitoring programs.Item Remote sensing of wet snow processes in a controlled laboratory environment(Montana State University - Bozeman, College of Engineering, 2022) Donahue, Christopher Paul; Chairperson, Graduate Committee: Kevin Hammonds; This is a manuscript style paper that includes co-authored chapters.Water flow through snow, due to snowmelt or rain-on-snow events, is a heterogeneous process that has implications for snowmelt timing and magnitude, snow metamorphism, albedo evolution, and avalanche hazard. Remote sensing technologies, ranging from ground-based to satellite-borne scales, offer a non-destructive method for monitoring seasonal snowpacks, although there is no single technique that is ideal for monitoring snow. Wet snow, specifically, presents a challenge to both optical and radar remote sensing retrievals. The primary aim of this dissertation was to develop wet snow remote sensing methods from within a controlled laboratory environment, allowing for precise characterization of snow properties. Experiments were conducted by preparing laboratory snow samples of prescribed structures and monitoring them during and after melt using hyperspectral imaging and polarimetric radar. Snow properties were characterized using X-ray computed microtomography, a dielectric liquid water content sensor, and serial-section reconstructions. In addition to laboratory experiments, hyperspectral imaging snow property retrieval methods were developed and tested in the field during wet snow conditions at the ground-based scale. The primary outcomes from this work were three new remote sensing applications for monitoring wet snow processes. First, a new hyperspectral imaging method to map effective snow grain size was developed and used to quantify grain growth due to wet snow metamorphism. Second, the optimal radiative transfer mixing model to simulate wet snow reflectance was determined and used to map liquid water content in snow in 2- and 3- dimensions. Lastly, snow melt progression was monitored using continuous upward-looking polarimetric radar and it was found, by comparison to 3-dimensional liquid water content retrievals from hyperspectral imaging, that the cross-polarized radar signal was sensitive to the presence of preferential flow paths. The work presented here highlights the utility of using a multi-sensor fusion approach to snow remote sensing. Although these laboratory remote sensing experiments were at a small scale, the remote sensing instrument response to specific snow conditions directly translates to larger scales, which is valuable to support algorithm development for ground, airborne, and spaceborne remote sensing missions.Item Towards reduced-cost hyperspectral and multispectral image classification(Montana State University - Bozeman, College of Engineering, 2021) Morales Luna, Giorgio L.; Chairperson, Graduate Committee: John SheppardIn recent years, Hyperspectral Imaging systems (HSI) have become a powerful source for reliable data in applications such as remote sensing, agriculture, and biomedicine. However, the abundant spectral and spatial information of hyperspectral images makes them highly complex, which leads to the need for specialized Machine Learning algorithms to process and classify them. In that sense, the contribution of this thesis is multi-folded. We present a low-cost convolutional neural network designed for hyperspectral image classification called Hyper3DNet. Its architecture consists of two parts: a series of densely connected 3-D convolutions used as a feature extractor, and a series of 2-D separable convolutions used as a spatial encoder. We show that this design involves fewer trainable parameters compared to other approaches, yet without detriment to its performance. Furthermore, having observed that hyperspectral images benefit from methods to reduce the number of spectral bands while retaining the most useful information for a specific application, we present two novel hyperspectral dimensionality reduction techniques. First, we propose a filter-based method called Inter-Band Redundancy Analysis (IBRA) based on a collinearity analysis between a band and its neighbors. This analysis helps to remove redundant bands and dramatically reduces the search space. Second, we apply a wrapper-based approach called Greedy Spectral Selection (GSS) to the results of IBRA to select bands based on their information entropy values and train a compact Convolutional Neural Network to evaluate the performance of the current selection. We also propose a feature extraction framework that consists of two main steps: first, it reduces the total number of bands using IBRA; then, it can use any feature extraction method to obtain the desired number of feature channels. Finally, we use the original hyperspectral data cube to simulate the process of using actual filters in a multispectral imager. Experimental results show that our proposed Hyper3DNet architecture in conjunction with our dimensionality reduction techniques yields better classification results than the compared methods, producing more suitable results for a multispectral sensor design.Item Spatial patterns in soil depth and implications for offseason nitrogen dynamics in dryland wheat systems of central Montana(Montana State University - Bozeman, College of Agriculture, 2022) Fordyce, Simon Isaac; Co-chairs, Graduate Committee: Clain Jones and Craig Carr; Pat Carr, Clain Jones, Jed Eberly, Scott Powell, Adam Sigler and Stephanie Ewing were co-authors of the article, 'Exploring relationships between soil depth and multi-temporal spectral reflectance in a semi-arid agroecosystem: effects of spatial and temporal resolution' submitted to the journal 'Remote Sensing of environment' which is contained within this thesis.; Pat Carr, Clain Jones, Jed Eberly, Rob Payn, Adam Sigler and Stephanie Ewing were co-authors of the article, 'Spatiotemporal patterns of nitrogen mineralization in a dryland wheat system' submitted to the journal 'Agriculture, ecosystems, and environment' which is contained within this thesis.Shallow soils (< 50 cm) under dryland wheat (Triticum aestivum L.) production lose large amounts of inorganic nitrogen (N) to leaching. Crops grown in shallow soils may be more responsive to N fertilizer due to lower fertilizer recovery and suppressed mineralization, raising questions as to whether standard practices of N fertilizer rate determination can increase risks of leaching and groundwater contamination in these environments. Mineralized N can be a major nutritional supplement for wheat crops in dryland agroecosystems, so accurate estimates of mineralization inputs can have important economic and environmental implications. To assess the potential for suppressed N mineralization in shallow soils, we used spectral reflectance from up to three sensors (unmanned aerial vehicle, National Agricultural Imagery Program, and Sentinel 2) to spatially characterize soil depth on three fields in Central Montana (Chapter 2) and compared surface (0-20 cm) carbon and N cycling indices across soil depth classes (Chapter 3). Carbon dynamics were stable across depth classes while N mineralization was lower in the shallow class. Results confirm multispectral imagery as a valuable tool for non-destructively characterizing fine-scale spatial patterns in soil depth and corroborate previous findings of lower N mineralization in shallow soil environments. Given the potential for heightened fertilizer responsiveness due to lower mineralization in these environments, decision support systems for site-specific fertility management (e.g., variable rate fertilizer application) should assess the environmental consequences of leaching alongside the economic benefits of applied fertilizer rates which maximize responses of yield, quality and same-year net revenue.Item Effect of spectral band selection and bandwidth on weed detection in agricultural fields using hyperspectral remote sensing(Montana State University - Bozeman, College of Agriculture, 2017) Tittle, Samuel Bryant; Chairperson, Graduate Committee: Rick L. LawrencePresence of weeds in agricultural fields affects farmers' economic returns by increasing herbicide input. Application of herbicides traditionally consists of uniform application across fields, even though weed locations can be spatially variable within a field. The concept of spot spraying seeks to reduce farmers' costs and chemical inputs to the environment by only applying herbicides to infested areas. Current spot spraying technology relies on broad spectral bands with limited ability to differentiate weed species from crops. Hyperspectral remote sensing (many narrow, contiguous spectral bands) has been shown in previous research to successfully distinguish weeds from other vegetation. Hyperspectral sensor technology, however, might not currently be practical for on-tractor applications. The research objectives were to determine (1) the utility of using a limited number of narrow spectral bands as compared to a full set of hyperspectral bands and (2) the relative accuracy of narrow spectral bands compared to wider spectral bands. Answers to these objectives have the potential for improving on-tractor weed detection sensors. Reference data was provided by field observations of 224 weed infested and 304 uninfested locations within two winter wheat fields in Gallatin County, Montana, USA. Airborne hyperspectral data collected concurrently with the reference data provided 6-nm spectral bands that were used in varying combinations and artificially widened to address the research objectives. Band selection was compared using Euclidean, divergence, transformed divergence, and Jefferies-Matusita signature separability measures. Certain three and four narrow band combinations produced accuracies with no statistical difference from the full set of hyperspectral bands (based on kappa statistic analysis, alpha = 0.05). Bands that were artificially widened to 96 nm also showed no statistically significant difference from the use of 6-nm bands for both all bands and select band combinations. Results indicate the potential for bands that can differentiate weed species from crops and that the narrowest spectral bands available might not be necessary for accurate classification. Further research is needed to determine the robustness of this analysis, including whether a single set of spectral bands can be used effectively across multiple crop/weed systems, or whether band selection is site or system specific.Item Observations and modeling of plasma flows driven by solar flares(Montana State University - Bozeman, College of Letters & Science, 2016) Brannon, Sean Robert; Chairperson, Graduate Committee: Dana W. Longcope; Dana Longcope was a co-author of the article, 'Modeling properties of chromospheric evaporation driven by thermal conduction fronts from reconnection shocks' in the journal 'The astrophysical journal' which is contained within this thesis.; Dana W. Longcope and Jiong Qiu were co-authors of the article, 'Spectroscopic observations of evolving flare ribbon substructure suggesting origin in current sheet waves' in the journal 'The astrophysical journal' which is contained within this thesis.One of the fundamental statements that can be made about the solar atmosphere is that it is structured. This structuring is generally believed to be the result of both the arrangement of the magnetic field in the corona and the distribution of plasma along magnetic loops. The standard model of solar flares involves plasma transported into coronal loops via a process known as chromospheric evaporation, and the resulting evolution of the are loops is believed to be sensitive to the physical mechanism of energy input into the chromosphere by the are. We present here the results of three investigations into chromospheric plasma flows driven by solar are energy release and transport. First, we develop a 1-D hydrodynamic code to simulate the response of a simplified model chromosphere to energy input via thermal conduction from reconnection-driven shocks. We use the results from a set of simulations spanning a parameter space in both shock speed and chromospheric-to-coronal temperature ratio to infer power-law relationships between these quantities and observable evaporation properties. Second, we use imaging and spectral observations of a quasi-periodic oscillation of a are ribbon to determine the phase relationship between Doppler shifts of the ribbon plasma and the oscillation. The phase difference we find leads us to suggest an origin in a current sheet instability. Finally, we use imaging and spectral data of an on-disk are event and resulting are loop plasma flows to generally validate the standard picture of are loop evolution, including evaporation, cooling time, and draining downflows, and we use a simple free-fall model to produce the first direct comparison between observed and synthetic downflow spectra.