Theses and Dissertations at Montana State University (MSU)
Permanent URI for this collectionhttps://scholarworks.montana.edu/handle/1/733
Browse
7 results
Search Results
Item 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.Item Utilizing distributions of variable influence for feature selection in hyperspectral images(Montana State University - Bozeman, College of Engineering, 2019) Walton, Neil Stewart; Chairperson, Graduate Committee: John SheppardOptical sensing has been applied as an important tool in many different domains. Specifically, hyperspectral imaging has enjoyed success in a variety of tasks ranging from plant species classification to ripeness evaluation in produce. Although effective, hyperspectral imaging can be prohibitively expensive to deploy at scale. In the first half of this thesis, we develop a method to assist in designing a low-cost multispectral imager for produce monitoring by using a genetic algorithm (GA) that simultaneously selects a subset of informative wavelengths and identifies effective filter bandwidths for such an imager. Instead of selecting the single fittest member of the final population as our solution, we fit a univariate Gaussian mixture model to a histogram of the overall GA population, selecting the wavelengths associated with the peaks of the distributions as our solution. By evaluating the entire population, rather than a single solution, we are also able to specify filter bandwidths by calculating the standard deviations of the Gaussian distributions and computing the full-width at half-maximum values. In our experiments, we find that this novel histogram-based method for feature selection is effective when compared to both the standard GA and partial least squares discriminant analysis. In the second half of this thesis, we investigate how common feature selection frameworks such as feature ranking, forward selection, and backward elimination break down when faced with the multicollinearity present in hyperspectral data. We then propose two novel algorithms, Variable Importance for Distribution-based Feature Selection (VI-DFS) and Layer-wise Relevance Propagation for Distribution-based Feature Selection (LRP-DFS), that make use of variable importance and feature relevance, respectively. Both methods operate by fitting Gaussian mixture models to the plots of their respective scores over the input wavelengths and select the wavelengths associated with the peaks of each Gaussian component. In our experiments, we find that both novel methods outperform variable ranking, forward selection, and backward elimination and are competitive with the genetic algorithm over all datasets considered.Item Thermal biology of the lesser grain borer Rhyzopertha dominica (F.) (Bostrichidae) and the warehouse pirate bug Xylocoris flavipes (Reuter) (Anthocoridae)(Montana State University - Bozeman, College of Agriculture, 2002) Campbell, Tracy Lynn MummItem An analysis of the market structure for Montana barley and potential outlets(Montana State University - Bozeman, College of Agriculture, 1957) Fedje, Duane L.Item Marketing of malting and feed barley in Montana and in the United States(Montana State University - Bozeman, College of Agriculture, 1966) Vaughan, E. DeanItem Measurement costs and pricing methods in the retail produce market(Montana State University - Bozeman, College of Agriculture, 1999) Malishka, Peter; Chairperson, Graduate Committee: Randal R. Rucker.A persistent practice in the retail produce market is the mixed use of per unit and per pound pricing for bulk produce commodities. While per pound pricing explicitly prices the size dimension of the produce, per unit pricing (known in the industry as "by the each" pricing) is a form of average pricing whereby units differing in size and value are sold for the same price. When goods are average priced, opportunities exist for buyers to find units of exceptional value at the going price. Exploiting these opportunities requires buyers to measure and compare the values of individual units. Measurement of this kind often results in costly wealth transfers among buyers and between buyers and sellers. Profit maximization implies that sellers will avoid average pricing and its associated measurement costs whenever alternative pricing methods can be implemented at lower cost. This study examines the implications of measurement costs in the retail produce market, and develops predictions concerning the seller's decision to set an average price (price per each) or a price per pound. Logistic regression analysis is used to test the predictions on retail price data from major retailers in Bozeman, Montana. The results suggest that sellers choose between the two pricing methods in a manner that is consistent with the minimization of pre-sale measurement costs.Item Minimum-data analysis of ecosystem service supply with risk averse decision makers(Montana State University - Bozeman, College of Letters & Science, 2009) Smart, Francis Clayton; Chairperson, Graduate Committee: John Antle.There is a need for models that produce results that are both timely and sufficiently accurate to be useful to policy makers. The minimum-data approach of Antle and Valdivia (2006) responds to this need by supplying a spatially explicit first order approximation that models ecosystem supply by producers. However, producers in developing nations often are observed to deviate from simple expected profit maximization. Risk is one possible explanation for this divergence. This study builds upon the minimum-data approach by allowing for risk averse producer preferences. The study presents a framework for translating relative risk aversion measurements into the parameters needed for the mean-standard deviation utility function. This study utilizes experimental and econometric measurements of risk aversion by other researchers to parameterize the model. Historic weather data are used with crop yield models to simulate temporal variation in crop yields. The model is used to simulate the supply of carbon sequestration in Machakos, Kenya. At low levels of risk, producers behave in a manner consistent with risk neutrality. However as risks and risk aversion levels increase, there is an increasing divergence from the behavior implied by expected profit maximization. The effects of varying the structure of risk preferences were also examined. This study finds that, consistent with the results in a number of other studies, the level of risk aversion is generally a more important factor in simulated behavior than the structure of risk preferences. This study also examines the effects of increasing the spatial variation of returns. As the spatial variation of returns increases, the predicted producer behavior converges on a fifty percent rate of adoption of the carbon sequestering system, regardless of other parameters. Overall, this study finds that - at levels of risk aversion measured in similar populations in developing nations - the inclusion of risk aversion in the model provides an explanation for why the observed behavior of producers appears to diverge from expected profit maximization.