Scholarship & Research

Permanent URI for this communityhttps://scholarworks.montana.edu/handle/1/1

Browse

Search Results

Now showing 1 - 10 of 24

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.
Using sparse coding as a preprocessing technique for insect detection in pulsed LIDAR data
(Montana State University - Bozeman, College of Engineering, 2022) Zsidisin, Connor Reece; Chairperson, Graduate Committee: Brad Whitaker
This research proposes using sparse coding as a preprocessing technique on insect lidar based data. This preprocessing technique will be used in conjunction with the Adaptive Boosting (AdaBoost), Random UnderSampling Boosting (RUSBoost), and neural network algorithms to automatically detect insects. The project aims to increase the effectiveness of these algorithms by using new images created by sparse coding. The K-Singular Value Decomposition (KSVD) algorithm will be used to train a dictionary on images that contain the majority class (non-insects). This trained dictionary will be used along with Orthogonal Matching Pursuit (OMP) to reconstruct all lidar images. The difference between the original image and the reconstructed image will be taken and processed by the feature extraction function and then used to train and test the models. Using a complete and an overcomplete dictionary our results show that the algorithms are able to detect insects at a higher rate. Using an overcomplete dictionary we are able to classify 93.18% of insect containing images in the testing dataset. Using the complete dictionary we were able to maintain 99.70% of non-insect images while increasing the percentage of insects classified to 84.09%.
Diode-laser-based high spectral resolution LIDAR
(Montana State University - Bozeman, College of Engineering, 2021) Colberg, Luke Stewart; Chairperson, Graduate Committee: Kevin S. Repasky
This thesis describes the design, construction, and testing of a high spectral resolution lidar (HSRL) as a part of a combined HSRL and differential absorption lidar (DIAL) system. The combined HSRL and DIAL instrument is constructed using the MicroPulse DIAL (MPD) architecture and uses distributed Bragg reflector lasers. The MPD architecture is unique because it is eye-safe and cost-effective; therefore, it is ideal for creating a network of ground-based lidars. This instrument is designed for thermodynamic profiling of the lower troposphere. A network of these instruments would be helpful for wide-scale atmospheric monitoring for weather forecasting and climate science. The purpose of the HSRL is to retrieve the optical properties of aerosols in the lower troposphere. The HSRL uses the DIAL offline laser, which has a wavelength of 770.1085 nm, and a potassium vapor cell as the spectral filter. The data retrieved from the HSRL provides the aerosol backscatter coefficient and the backscatter ratio up to an altitude of 7 km during nighttime operation and 5 km during daytime operation. The time resolution for these measurements is 5 minutes, and the range resolution is 150 m. These aerosol optical properties are valuable for aerosol studies and climate modeling; aerosols introduce the most significant degree of uncertainty in modeling the heat flux of the atmosphere. Additionally, these aerosol optical properties can be used to find the planetary boundary layer height (PBLH). The planetary boundary layer controls the exchange of heat, water vapor, aerosols, and momentum between the surface and the atmosphere. It has been demonstrated that the PBLH strongly affects turbulent mixing, convective transport, and cloud entrainment, which makes the PBLH an important parameter for weather forecasting and climate modeling. Despite its significance in atmospheric science, there is no standard method for defining the PBLH. A retrieval method for finding the daytime PBLH using HSRL data is proposed, and data comparisons to radiosonde PBLH retrievals are provided. The algorithm shows a good agreement with the radiosonde retrievals for conditions with a well-behaved boundary layer.
Machine learning pipeline for rare-event detection in synthetic-aperture radar and LIDAR data
(Montana State University - Bozeman, College of Engineering, 2021) Scofield, Trey Palmer; Chairperson, Graduate Committee: Brad Whitaker
In this work, we develop a machine learning pipeline to autonomously classify synthetic aperture radar (SAR) and lidar data in rare-event, remote sensing applications. Here, we are predicting the presence of volcanoes on the surface of Venus, fish in Yellowstone Lake, and select marine-life in the Gulf of Mexico. Given the efficiency of collecting SAR images in space and airborne lidar geographical surveys, the size of the datasets are immense. Immense training data is desirable for machine learning models; however, a large majority of the data we are using do not contain volcanoes or fish, respectively. Thus, the machine learning models must be formulated in such a way to place a high emphasis on the minority, target classes. The developed pipeline includes data preprocessing, unsupervised clustering, feature extraction, and classification. For each collection of data, sub-images are initially fed through the pipeline to capture fine detail characteristics until they are mapped back to their original image to identify overall region behavior and the location of the target class(es). For both sub-images and original images, results were quantified and the most effective algorithm combinations and parameters were assigned. In this analysis, we determined the classification results are not sufficient enough to propel a completely autonomous system, rather, some manual observing of the data will need to be performed. Nonetheless, the pipeline serves as an effective tool to reduce costs associated with electronic storage and transmission of the data, as well as human labor in manually inspecting the data. It does this by removing a majority of the unimportant, non-target data in some cases while successfully retaining a high percentage of the important images.
Results of a micro pulse differential absorption LIDAR for temperature profiling and analysis code
(Montana State University - Bozeman, College of Engineering, 2021) Cruikshank, Owen Daniel; Chairperson, Graduate Committee: Kevin S. Repasky
Thermodynamic profiling of the lower troposphere is necessary for the study of weather and climate. The micropulse DIAL (differential absorption lidar), or MPD, presented here is designed to fill the need. The MPD is eye-safe and can run autonomously for continuous measurements compared to technologies with similar measurement capabilities like Raman lidar. Using a temperature-sensitive absorption line of O 2, the MPD system can measure the absorption of O 2 in the lower troposphere as a function of range and convert that measurement to temperature as a function of range. This process relies on a perturbative correction to the absorption retrieval to account for the fact that the O 2 absorption spectral linewidth is similar to the molecular Rayleigh scattering linewidth. An ancillary measurement of the ratio of aerosol backscatter to molecular backscatter is required for the correction. The integrated high spectral resolution lidar (HSRL) uses a heated potassium vapor notch filter to make the aerosol-to-molecular ratio measurement. An analysis program in MATLAB was written to take in raw lidar data and produce a temperature product of range and time. Results presented from a campaign at the Atmospheric Radiation Measurements program Southern Great Plains site in Oklahoma in spring 2019 show temperature comparisons with radiosonde measurements with a mean difference between radiosonde and MPD measurements of -1.1K and a standard deviation of 2.7 K. Further results from an instrument on the Montana State University campus in Bozeman and at the National Center for Atmospheric Research in Boulder, Colorado have shown that the MPD instrument can produce measurements autonomously for periods of weeks to months.
Calibration and characterization of a VNIR hyperspectral imager for produce monitoring
(Montana State University - Bozeman, College of Engineering, 2020) Logan, Riley Donovan; Chairperson, Graduate Committee: Joseph A. Shaw; Joseph A. Shaw was a co-author of the article, 'Measuring the polarization response of a VNIR hyperspectral imager' in the journal 'SPIE proceedings' which is contained within this thesis.; Bryan Scherrer, Jacob Senecal, Neil S. Walton, Amy Peerlinck, John W. Sheppard, and Joseph A. Shaw were co-authors of the article, 'Hyperspectral imaging and machine learning for monitoring produce ripeness' in the journal 'SPIE proceedings' which is contained within this thesis.
Hyperspectral imaging is a powerful remote sensing tool capable of capturing rich spectral and spatial information. Although the origins of hyperspectral imaging are in terrestrial remote sensing, new applications are emerging rapidly. Owing to its non-destructive nature, hyperspectral imaging has become a useful tool for monitoring produce ripeness. This paper describes the process of characterizing and calibrating a visible near-infrared (VNIR) hyperspectral imager for obtaining accurate images of produce to be used in machine learning algorithms for analysis. In this work, many calibrations and characterization are outlined, including: a radiance calibration, the process of calculating reflectance, pixel uniformity and image stability testing, spectral characterization, illumination source analysis, and measurement of the polarization response. The images obtained by the calibrated hyperspectral imager were converted to reflectance across a spectral range of 387.12 nm to 1023.5 nm, with a spectral resolution of 2.12 nm. A convolutional neural network was used to perform age classification for Yukon Gold potatoes, bananas, and green peppers. Additionally, a genetic algorithm was used to determine the wavelengths carrying the most useful information for age classification. Experiments were run using red green blue (RGB) images, full-spectrum hyperspectral images, and the wavelengths selected by the genetic algorithm feature selection method. Preliminary data from these analyses show promising results at accurately classifying produce age. The genetic algorithm feature selection method is being used to develop a low-cost multispectral imager for use in monitoring produce in grocery stores.
Characterization of a division-of-focal-plane polarization imager
(Montana State University - Bozeman, College of Engineering, 2020) Syed, Musaddeque Anwar Al Abedin; Chairperson, Graduate Committee: Joseph A. Shaw
Polarization is a fundamental property of light that can be detected with polarization-sensitive instruments. Imaging polarimetry has an immensely wide range of applications, and while much has been accomplished in recent years, there is still a need for sensor systems with improved accuracy, precision, and stability. This paper presents the optical characterization of a commercial division-of-focal plane (DoFP) polarization imager, in an effort to evaluate its performance as a promising instrument in the application of ground-based cloud thermodynamic phase detection. Radiometric characterization values were well within the acceptable region, but the polarimetric contrast was in the range of 20-30, much lower than expected, which may be a result of the broadband measurements being impaired by poor polarizer performance at the blue end of the spectrum. Later, a narrowband polarimetric measurement at 532 + or - 5 nm produced a much enhanced result, with polarimetric contrast in the higher 300s, making the imager a viable option for many remote sensing applications. Also, all-sky imaging of clear daytime sky and its analysis of degree of linear polarization (DoLP) showed encouraging result.
Combining spectral and polarimetric methods to classify cloud thermodynamic phase
(Montana State University - Bozeman, College of Engineering, 2019) Tauc, Martin Jan; Chairperson, Graduate Committee: Joseph A. Shaw; David W. Riesland, Laura M. Eshelman, Wataru Nakagawa and Joseph A. Shaw were co-authors of the article, 'Radiance ratios for CTP discrimination' submitted to the journal 'Journal of applied remote sensing' which is contained within this thesis.; Wataru Nakagawa and Joseph A. Shaw were co-authors of the article, 'The SWIR three-channel polarimeter for cloud thermodynamic phase detection' in the journal 'Optical engineering' which is contained within this thesis.
Cloud thermodynamic phase--whether a cloud is composed of spherical water droplets or polyhedral ice crystals--is an important parameter for optical communication with space-based instruments, remote sensing of the atmosphere, and, perhaps most importantly, understanding weather and climate. Although some methods exist to detect the phase of clouds, there is still a need for passive remote sensing of cloud thermodynamic phase due to its low-cost, scalability, and ease of use. Two methods for cloud thermodynamic phase classification employ spectral radiance ratios in the short-wave infrared, and the S 1 Stokes parameter, a polarimetric quantity. In this dissertation, the combination of the two methods is realized in an instrument called the short-wave infrared three-channel polarimeter. The coalescence of radiance ratios in the short-wave infrared and polarization channels oriented parallel and perpendicular to the scattering plane provides better classification of cloud phase than either method independently. Despite the improvement, the low-cost system suffered from hardware and software limitations, which caused an increase in noise and polarimetric artifacts. These errors are analyzed and a subset of low-noise data shows even better classification ability. All together, the results attained from the deployment of the polarimeter in early 2019 showed promise that the combination of the two methods is an improvement over past techniques.
Weed and crop discrimination with hyperspectral imaging and machine learning
(Montana State University - Bozeman, College of Engineering, 2019) Scherrer, Bryan Joseph; Chairperson, Graduate Committee: Joseph A. Shaw
Herbicide-resistant weed biotypes are spreading across crop fields nationally and internationally and mapping them with traditional crop science methods - cloning plants and testing their resistance levels in a lab - are costly and time consuming. A segment of the field of precision agriculture is being developed to accurately and quickly map the location of herbicide-resistant and herbicide-susceptible weeds using advanced optics and computer algorithms. In our study, we collected hundreds of thousands of spectra of herbicide-resistant and herbicide-susceptible biotypes of the weeds kochia, mare's tail and lamb's quarter and of crops including barley, corn, dry pea, garbanzo, lentils, pinto bean, safflower, sugar beet at the Southern Agricultural Research Center in Huntley, Montana using a hyperspectral imager. Plants were imaged in a controlled greenhouse setting as well as in crop fields using ground-based and drone-based imaging platforms. The spectra were differentiated from one another using a feedforward neural network machine learning algorithm. Classification accuracies depended on what plants were imaged, the age of the plants and lighting conditions of the experiment. They ranged from 77% to 99% for spectra acquired on our ground-based imaging platform and from 25% to 79% on our drone- based platform.
A study of atmospheric polarization in unique scattering conditions at twilight, during a solar eclipse, and for cloud phase retrievals using all-sky polarization imaging
(Montana State University - Bozeman, College of Engineering, 2018) Eschelman, Laura Marie; Chairperson, Graduate Committee: Joseph A. Shaw
Polarization is a fundamental property of light that can be detected with polarization-sensitive instruments for many remote sensing applications. To quantitatively interpret the remote sensing data, an understanding how naturally occurring polarization depends on wavelength and environmental parameters is needed. The most obvious source of naturally occurring polarization is atmospheric scattering. For a clear-sky environment, Rayleigh scattering dominates, resulting from scattering by atmospheric gas molecules that are much smaller than the optical wavelength, and a distinct all-sky polarization pattern exists. A band of maximum degree of linear polarization can be observed 90? from the sun with polarization vectors orientated perpendicular to the scattering plane (i.e. the plane containing the incident and scattered light). However, aerosols, clouds, and underlying surface reflectance can alter the observed sky polarization. Military, environmental, and navigational applications exploit the sky polarization pattern to detect objects, retrieve aerosol and cloud properties, and to find compass headings based on the sky polarization pattern. Sky polarization is also being used to calibrate the polarization response of large telescopes. It is important to understand how partially polarized skylight can vary with environmental factors, as well as with wavelength and solar position, so that polarization measurements can be interpreted correctly. The direction of polarization when aligned to a specific reference frame can provide additional information beyond the basic polarization pattern. This dissertation expands the current knowledge of skylight polarization by validating radiative transfer simulations in the shortwave infrared, by reporting the first-ever retrievals of cloud thermodynamic phase from all-sky polarization images using the Stokes S1 parameter referenced in the scattering plane, and by quantifying how partially polarized skylight varied under unique scattering conditions during the 2017 solar eclipse. In order to accurately predict cloud thermodynamic phase and to analyze the temporal distribution of skylight during a total solar eclipse, a physics-based understanding of the Stokes parameters and angle of polarization (AoP) with respect to the instrument, scattering, and solar principal planes was also developed. Through each experiment, two underlying threads were observed. First, in order to accurately interpret results, environmental parameters needed to be characterized. Second, when rotated into a specific reference frame, the Stokes parameters and AoP can be utilized differently and provide unique insights when analyzing all-sky polarization data.

Scholarship & Research

Browse

Filters

Settings

Sort By

Results per page

Search Results