DalSpace Institutional Repository

DalSpace is a digital service that collects, preserves, and distributes digital material produced by the Dalhousie community.

To learn about content guidelines, policies, and how to deposit, view the Help documents.
Contact us to get started submitting content to DalSpace at dalspace@dal.ca

Communities in DSpace

Select a community to browse its collections.

Now showing 1 - 10 of 19

Recent Submissions

Access status: Embargo ,
Development of full-scale machine vision scheme of multi-camera systems on boom sprayer for real-time management of Colorado Potato Beetles
(2026-04-28) Hassan, Imran; Not Applicable; Doctor of Philosophy; Faculty of Agriculture; Not Applicable; Mark Eramian; Not Applicable; Qamar Zaman; Khalil Al-Mughrabi; Ahmad Al-Mallahi
Precision agriculture demands site-specific pest management to reduce chemical inputs and environmental harm, yet real-time detection of small insect pests from mobile platforms re-mains challenging. The Colorado potato beetle (CPB), measuring less than 10 mm in body length, is among the most economically destructive pests in potato production globally and has developed resistance across all major insecticide classes. This research addresses the gap between laboratory-validated detection algorithms and field-deployable systems by developing, integrating, and evaluating a machine vision scheme for real-time CPB detection under commercial operating conditions. A field dataset was constructed from imagery captured using static handheld and moving sprayer-mounted cameras across two growing seasons, with bounding-box annotation under variable lighting, diverse canopy structures, and natural backgrounds. An image-cropping-based pre-processing technique was developed to preserve fine-scale morphological features suppressed by conventional input resizing, directly addressing the tiny object detection challenge in high-resolution agricultural imagery. Three deep learning architecture — YOLOv5, YOLOv7, and Faster R-CNN — were trained and benchmarked; YOLOv5 was selected for deployment on its favorable accuracy–latency trade-off, achieving 78 % accuracy on independent test data while sustaining real-time throughput. A modular software pipeline was engineered for concurrent multi-camera stream processing, combining TensorRT-accelerated inference with deterministic spray command generation and network-based communication with sprayer control systems. End-to-end latency measured approximately 120 milliseconds across four simultaneous camera feeds. Full-scale machine vision scheme deployment was implemented on a Case IH Patriot 3240 boom sprayer with eight machine vision systems across a 30-meter boom, each handling three to four downward-facing cameras. Engineering challenges resolved included distributed power supply, environmental enclosure de-sign, boom folding compatibility, and synchronized multi-unit operation under field vibration. Field validation at a commercial spraying speed of 2.7 m/s confirmed stable real-time operation, with detection accuracy exceeding 80 %, precision near 90 %, and recall approaching 90 %. Furthermore, a novel five-crop dataset expansion technique increased effective training data fivefold while preserving object scale and introducing natural background diversity; benchmarking with state-of-the-art YOLOv11 architecture demonstrated mean average precision exceeding 85 % at standard IoU thresholds. This research establishes a validated framework for insect-specific precision spraying, with contributions in tiny object detection, real-time multi-camera processing, and mobile agricultural system integration.
Access status: Open Access ,
Multimodal Representation Learning for Mental Health: Transfer Learning, PEFT, and Contrastive Learning
(2026-04-28) Naderi Khorshidi, Habibeh; Not Applicable; Doctor of Philosophy; Faculty of Computer Science; Not Applicable; Dr. Diana Inkpen; Not Applicable; Dr. Evangelos Milios; Dr. Sageev Oore; Dr. Stan Matwin; Dr. Frank Rudzicz
This thesis addresses the challenge of learning effective multimodal representations in low-resource clinical settings by leveraging implicit supervision from paired modalities and introducing adaptive fusion mechanisms based on Mixture-of-Experts (MoE). Focusing on mental health assessment, it utilizes clinical audio interviews from parents and children to jointly model speech and language signals for predicting emotion, sentiment, and higher-level psychological and cognitive outcomes under conditions of data scarcity and class imbalance. In collaboration with the Department of Psychiatry at Dalhousie University, we curate a dataset of parent-child clinical interviews annotated at both segment and document levels, including fine-grained labels for emotion recognition, sentiment analysis, and criticism detection, as well as higher-level diagnostic outcomes such as ADHD, depression, and bipolar disorder. The thesis systematically investigates transfer learning, parameter-efficient fine-tuning (PEFT), and contrastive audio-text learning for multimodal representation learning. We first conduct a comprehensive study of transfer learning using 29 pre-trained language models and 22 audio models, evaluating their effectiveness under various adaptation strategies and analyzing the impact of limited and imbalanced data. Building on this, we explore PEFT as a more efficient alternative, applying six techniques, LoRA, AdaLoRA, OFT, LoHa, LoKr, and IA3, across multiple models. PEFT consistently outperforms conventional transfer learning, yielding improvements of +1.41\% and +0.8\% in average AUC on parent and offspring data, respectively. Finally, we propose a three-stage contrastive learning framework for multimodal representation learning. The approach combines unimodal encoding, cross-modal contrastive alignment, and downstream task learning with a MoE architecture that adaptively integrates modality-specific information. Across all evaluated tasks, the proposed frameworks consistently outperform strong unimodal and multimodal baselines, demonstrating the effectiveness of adaptive multimodal learning for mental health assessment in low-resource settings.
Access status: Open Access ,
THE APPLICATION OF PALEOLIMNOLOGY TO INFORM FRESHWATER MANAGEMENT: THE CASE OF URBAN LAKES IN HALIFAX REGIONAL MUNICIPALITY, NOVA SCOTIA, CANADA
(2026-04-28) Hipwell, Kathleen; Not Applicable; Master of Environmental Studies; School for Resource & Environmental Studies; Not Applicable; Dr. Josh Kurek; Yes; Dr. Alana Westwood; Dr. Andrew S. Medeiros
Watershed-scale stress from urbanization can negatively impact ecosystem services communities rely on. Yet, a lack of historical context for individual-lake conditions impedes management planning that promotes evidence-based restoration/remediation targets and the efficient allocation of resources. Halifax Regional Municipality (HRM) Nova Scotia, Canada, has hundreds of lakes that are increasingly influenced by residential development, and most lakes in the urban/suburban core are developed to some degree. Here, I assessed surface sediment subfossil chironomid assemblages (Diptera: Chironomidae) from 40 HRM lakes and used a development index to determine if there is a relationship between assemblage structure and watershed development level. Chironomid assemblages were analyzed to understand the broader context of urban influence and to identify lake-specific environmental stressors that influence assemblage structure with water quality parameters for a subset of 29 lakes. Through the additional biostratigraphic analysis of three lakes we used paleolimnology to fill the gap in long-term monitoring data and to infer the influence of watershed development on urban lakes (Chocolate Lake and Settle Lake lie in residentially developed watersheds, and Spider Lake is in an undeveloped watershed). I found that chironomid assemblages followed a productivity gradient with development level, and that conductivity and ions were the primary drivers explaining variation in surface sediment assemblages. In the two lakes in developed watersheds, shifts in chironomid assemblages over time reflected documented patterns of human disturbance and indicate increased productivity. In contrast, the lake in an undeveloped watershed has remained oligotrophic throughout its history. Our results could facilitate informed decisions related to restoration/remediation targets and assist in the planning and management of freshwater resources.
Access status: Open Access ,
The Marine Predation, Migration, and Habitat Use of Inner Bay of Fundy Atlantic Salmon (Salmo salar) Kelts
(2026-04-23) Brown, Caliyena; Not Applicable; Master of Science; Department of Biology; Not Applicable; Tommi Linnansaari; Not Applicable; Marc Trudel; Sarah Tuziak; Sarah Fortune; Robert Lennox
Atlantic salmon (Salmo salar) populations have declined across much of their range, and the inner Bay of Fundy (iBoF) population has experienced one of the most severe collapses in North America. Despite supplementation-based recovery efforts restoring juvenile freshwater abundance, the number of adults returning to spawning rivers remains critically low and repeat spawners are largely absent, indicating that poor marine survival is hindering population recovery. Using temperature-sensing acoustic transmitters, this thesis examines the marine predation, migration, and habitat use of captive-reared iBoF Atlantic salmon kelts (post-spawned adults) to better understand the marine ecology of this critical life stage. Chapter 2 evaluates the role of predation in kelt marine mortality. Predation was observed in ~ 37% of iBoF salmon kelts detected in the Bay of Fundy (BoF). Neither fish size (total length; m) nor fish sex significantly influenced predation risk, but distinct spatial and temporal patterns of predation were evident, with events occurring predominantly during late spring and summer in the outer BoF. Thermal signatures indicated that most predation events were attributable to mesothermic predators, with limited evidence of marine mammal predation. Chapter 3 characterizes marine migration and seasonal habitat use of iBoF salmon kelts within the BoF. Seasonal modelling revealed increased kelt presence in the middle and outer BoF during late spring and summer, with spatial use shifting across months and concentrating in regions previously identified as important marine habitat. Kelts exhibited variable movement patterns, including both prolonged residency within the BoF and streamlined exits into the Gulf of Maine. About one third of the kelts exited the Bay at least once, primarily in May and June. Collectively, these findings provide contemporary evidence that predation is a prominent driver of marine mortality for iBoF salmon kelts, and that both survival and marine movements vary across space and time. This work directly informs recovery planning for this endangered population, including improving understanding of factors influencing marine survival, refinement of hatchery release strategies, and progress towards identifying important marine habitat for iBoF salmon kelts.
Access status: Embargo ,
Genomic Characterization of Five Isolates of Blastocystis Subtype 7
(2026-04-27) Seaton, Gregory; No; Doctor of Philosophy; Department of Biochemistry & Molecular Biology; Not Applicable; Dr. Laura Wegener Parfrey; No; Dr. Claudio H. Slamovits; Dr. James M. Kramer; Dr. Morgan Langille; Dr. Andrew Roger
Blastocystis spp. are common, genetically diverse anaerobic colonizers of the gastrointestinal tracts of both humans and animals. Although they are traditionally regarded as parasites, their precise role in host health remains unclear, in part because cryptic genetic diversity is masked by morphological similarity among divergent lineages. Subtype 7 (ST7) contains several isolates (C, H, E, G and B) that have been shown to exhibit distinct phenotypic differences in experimental settings, suggesting genetic diversity within the subtype. This thesis presents a comparative genomic analysis of those five ST7 isolates, with a focus on genetic differences between the isolates. Genomes of four of the five isolates presented here are newly sequenced using a long-read-based assembly approach resulting in new genomic assemblies that are >98% complete and range from 19.8 (ST7C) to 20.2 (ST7G) Mbp in size and were found to be highly syntenic. Short-read data suggests these Blastocystis ST7 isolates are haploid. For two of the newly sequenced genomes, ST7C and ST7G, the assembly was complete enough to capture 8 and 9 complete chromosomes respectively. Through comparisons of shared syntenic scaffolds amongst isolates, 16 nuclear chromosomes could be hypothetically reconstructed in ST7C. Comparative analyses revealed clear evidence of genomic rearrangements in ST7G and ST7H relative to ST7C. Manual curation was used to create a high-quality gene annotation for ST7C that was propagated to the other genomes yielding 8625-8896 predicted genes per genome. The predicted gene set for ST7B contained 2593 more genes than a previously published genome analysis for this isolate. A pangenome analysis of the five isolates was conducted and a small (228) accessory genome was identified compared to a far larger core gene set (6198) shared by all the isolates. An abundance of endogenous viral elements belonging to the Midsized Eukaryotic Linear dsDNA (MELD) virus class were found and comparative analyses suggest that the MELD viruses first invaded the common ancestor of ST7 and ST6, and then greatly proliferated in the ST7 genomes. The analyses presented here offers insight into the genomic structure and differences between these Blastocystis ST7 isolates.
Access status: Open Access ,
Seismic Retrofit of Dry Precast Concrete Beam To Column Joint Using Shape Memory Alloy
(2026-04-27) Chen, DeZhen; Not Applicable; Master of Applied Science; Department of Civil and Resource Engineering; Not Applicable; N/A; Not Applicable; Pedram Sadeghian; Zoheir Farhat; Fadi Oudah
Superelastic shape memory alloy (SMA) is regarded as an ideal material for enhancing the seismic capacity of structures due to its superelasticity and self-centering properties. This research introduces and validates a new concept for the structural upgrade of seismically deficient precast concrete structures using SMA. The proposed approach establishes partial moment transfer at the beam-column joint, enabling the frame to partially resist lateral loads alongside the building's shear core. To achieve this, a novel device—termed the SMA Partial Moment Transfer (SMA-PMT) device—is developed and validated through experimental testing and analytical modeling using hybrid simulation. The research was executed in three stages, including model development, the experimental joint design and the experimental test. The discussion summarizes the behavior of the experimental joint specimens, focusing on joint moment capacity (yield moment), energy dissipation capabilities, and joint damage assessments. Recommendations for future research are also provided.