Institutional Repository

U n i v e r s i t y of P e r a d e n i y a

The Institutional Repository of the University of Peradeniya is the University's digital gateway to scholarly knowledge and research excellence. Established as the next phase of the Digital Library initiative launched by the University Library in 2011, it preserves, showcases, and provides open access to the intellectual output of the University community.

Through a rich collection of research articles, conference proceedings, theses, dissertations, and other scholarly works, the repository enhances the visibility, accessibility, and long-term preservation of knowledge, extending the global reach and impact of the University of Peradeniya’s academic and research contributions.

Photo by @ITD,Main Library

Communities in DSpace

Select a community to browse its collections.

Now showing 1 - 5 of 5

Recent Submissions

An integrated computer vision framework for structural defect identification, quantification and multi-parameter severity assessment in reinforced concrete beams
(University of Peradeniya, Sri Lanka, 2026-08-20) Bandara, U. L. N. T.; Ariyaratne, H. M. P. B.; Jayasinghe, J. A. S. C.; Jayasinghe, U.
Current computer vision applications in Structural Health Monitoring (SHM) predominantly focus on binary crack detection, offering limited support for engineering condition assessment, prioritization of repairs, and maintenance decision-making. Existing studies have demonstrated reliable crack quantification; however, multi-parameter severity assessment frameworks that convert measured crack characteristics into actionable condition ratings remain comparatively underexplored. To address this gap, this study developed a comprehensive image-based framework integrating automated structural defect identification and precise geometric quantification directly with a robust severity assessment model tailored for Reinforced Concrete (RC) beams. A benchmark dataset of surface cracks exclusively in RC beams was constructed using images captured under realistic field and laboratory conditions, intentionally including varying illumination, texture, shadows, staining, and surface deterioration commonly encountered during inspections. A total of 300 high-resolution images were collected, with 100 images manually annotated for semantic segmentation. The dataset was divided into training, validation, and testing subsets of 70%, 20%, and 10%, respectively. The finalized dataset includes precise crack masks and image-derived parameters such as crack width, length, orientation, location, density, and branching characteristics. Semantic segmentation and advanced image-processing techniques were utilized to automatically extract these physical parameters, prioritizing critical serviceability thresholds and the spatial distribution of cracks. Additionally, image preprocessing procedures, including contrast enhancement and noise reduction, were successfully evaluated to improve visibility under challenging conditions. The quantified crack characteristics were subsequently integrated into a novel rule-based severity assessment framework informed by established engineering standards and structural judgement. By evaluating multiple interrelated parameters rather than relying on crack width alone, the framework successfully executed a severity assessment that classified the affected RC beams into Minor, Moderate, and Critical severity levels. Results demonstrated that integrating semantic segmentation with multi-parameter analysis successfully extracted geometric features capable of driving these accurate, standardized severity classifications. Model outputs were thoroughly benchmarked against manual measurements and engineer-assigned condition ratings, confirming high classification reliability. Ultimately, this framework successfully converted visual defect observations into actionable engineering ratings, significantly reducing reliance on subjective manual interpretation while providing a scalable methodology for rapid inspection, maintenance prioritization, and long-term monitoring of ageing RC beam infrastructure.
Accuracy comparison of machine learning and statistical methods for daily rainfall estimation in the Kelani River Basin, Sri Lanka
(University of Peradeniya, Sri Lanka, 2026-08-20) Kamburugamuwa, S. W.; De Silva, M. M. G. T.; Herath, H. M. D. K.; Neluwala, N. G. P. B.
Accurate estimation of daily rainfall at local scales remains a major challenge in hydrology due to the highly nonlinear and spatially variable nature of precipitation. For flood forecasting and water resource management, accurate rainfall prediction is essential. This study evaluates the performance of machine learning and conventional techniques for estimating daily rainfall at the Hanwella station, located within the Kelani River Basin, based on rainfall data from four neighbouring stations. An Extreme Gradient Boosting (XGBoost) model and a Long Short-Term Memory (LSTM) model were developed using multi-station rainfall data from 2020 to 2025, while spatial interpolation techniques such as Inverse Distance Weighting (IDW) and the Normal Ratio method were used as reference techniques. The LSTM model was trained using 3-day temporal sequences with various loss functions (Mean Absolute Error (MAE), Mean Squared Error (MSE), Huber loss, Mean Squared Logarithmic Error (MSLE), and Log-Cosh loss). The performance of all four models was assessed using Root Mean Square Error (RMSE), MAE, and Nash–Sutcliffe Efficiency (NSE). These evaluation metrics were calculated separately for the validation and testing datasets to provide a comprehensive comparison of model performance. Among the evaluated methods, the IDW approach (p = 2) exhibits the highest performance, achieving an RMSE of 18.32 mm/day and an NSE of 0.63, indicating a moderate level of predictive accuracy, whereas the remaining methods demonstrate comparatively lower predictive accuracy. XGBoost demonstrated competitive performance on the structured tabular dataset, followed by the Normal Ratio method, while LSTM showed the lowest overall accuracy. This behaviour may be attributed to the dominance of spatial relationships over temporal dynamics, thereby limiting the effectiveness of sequence-based learning. These outcomes imply that model configuration, feature representation, and data properties significantly impact model performance. The LSTM model has difficulty predicting peak rainfall events, even if it was able to capture overall rainfall variability. XGBoost shows superior performance on the structured dataset compared to LSTM, but both methods can be further improved. This study helps to make more accurate flood predictions and water resource management by offering useful insights for enhancing rainfall estimation techniques.
Accuracy comparison of lightweight vision-language models for multimodal fake news detection
(University of Peradeniya, Sri Lanka, 2026-08-20) Ekanayake, E. M. C. P.; Herath, D.
Fake news shared on digital platforms increasingly combines textual and visual content to spread misleading information. The Indian Fake News Dataset (IFND) is a multimodal benchmark for fake news detection in the Indian news domain. The use of lightweight vision-language models for multimodal fake news detection on the IFND dataset has not yet been explored. We focus on lightweight vision-language models because fake news detection on social media platforms requires models that can process large amounts of multimodal content efficiently and with reasonable computing cost. In this work, we conduct a comparative analysis of lightweight vision-language models for binary fake news detection on IFND. Each sample is represented as a caption-image pair, and the model predicts whether the multimodal content is true or fake. The study is conducted on a subset of IFND containing 25,224 samples. Our experiments compare the performance of six lightweight vision-language models on this benchmark. Among the evaluated models, InstructBLIP achieved the best test accuracy of 72%, while Qwen-VL showed the lowest test accuracy of 33%. These results show that lightweight vision-language models can be applied for multimodal fake news detection, although their performance differs considerably across model architectures. Overall, this work presents a benchmark of lightweight vision-language models on IFND and highlights their potential for practical multimodal misinformation detection.
A wireless soft PLC architecture for scalable and cost-efficient intelligent automation platform for precision agriculture
(University of Peradeniya, Sri Lanka, 2026-08-20) Siriwardena, S. J.; Bandara, A. S.; Ekanayake, J. B.; Samaranayake, B. G. L. T.
Intelligent automation in precision agriculture demands scalable, adaptive, and cost-effective control architectures capable of reliable operation across distributed environments. Conventional Programmable Logic Controllers (PLCs), while deterministic and robust, are limited by high cost, hardware dependency, and poor scalability in geographically dispersed applications. This work proposes a software-defined Soft Programmable Logic Controller (Soft-PLC) architecture that virtualizes traditional PLC functionality while maintaining deterministic execution and hierarchical organization. A unified cyber-physical framework is introduced using a single long-range LoRa communication network to support sensing, closed-loop control, and actuation over distances up to 1 km. Wireless sensor nodes, each powered by a dedicated solar panel with a backup battery pack, measure nitrogen, phosphorus, potassium, pH, temperature, humidity, and electrical conductivity. Data are transmitted to a centralized control hub at intervals ranging from 30 seconds to 5 minutes, depending on agronomic requirements. The unified bidirectional LoRa infrastructure replaces the multiple communication gateways and multi-network infrastructure of conventional PLC-based precision agriculture systems, reducing system complexity and communication hardware by 30– 50%. Control logic is executed on a Raspberry Pi-based hub using a deterministic scan-cycle paradigm (input acquisition, logic execution, output update) aligned with IEC 61131-3 standards, achieving local controller scan times of 40–80 ms. End-to-end communication performance is evaluated separately through LoRa network latency and packet loss. The system follows ISA-95 layered design principles to ensure structured interaction between field devices, control processes, and supervisory systems, supporting integration with existing industrial automation frameworks. The key contribution is the integration of Soft-PLC virtualization, unified LoRa communication, and ISA-95 architecture into a scalable, hardware-independent automation platform for precision agriculture. Field validation over 72 hours in an open agriculture field with node–hub distances from 200 m to 1 km demonstrates the feasibility and stable short-term operation of the proposed architecture, with an average communication latency of 180 ms (range: 150–250 ms) and a packet loss rate of 2.3% (below 5% across all nodes). The system achieves a 40–50% cost reduction compared to conventional PLC-based precision agriculture setups while maintaining comparable control performance, providing a scalable automation and communication platform capable of supporting future AI-driven precision agriculture applications.
A unified comparative analysis of attention mechanisms for pneumonia detection in chest X-ray images
(University of Peradeniya, Sri Lanka, 2026-08-20) Selvarasa, T.; Ramanan, A.
Lower respiratory tract infections continue to represent a significant global health burden, particularly among children in low-resource settings. Limited access to expert radiological interpretation further increases the need for reliable automated diagnostic systems using chest X-ray (CXR) imaging. This study presents a systematic and controlled comparative analysis of three widely used attention mechanisms: Squeeze-and-Excitation (SE), Spatial Attention (SA), and Convolutional Block Attention Module (CBAM) within a unified deep learning framework for pediatric pneumonia detection. Unlike existing studies, many approaches report results using inconsistent datasets, preprocessing methods, and training strategies, making fair comparison difficult. To address this limitation, the present study ensures fairness and reproducibility by integrating each attention mechanism into the same ResNet50 backbone architecture and training all models under identical experimental conditions. The study utilises a publicly available Chest X-ray dataset (https://tinyurl.com/kaggle-CXR) comprising 5,856 pediatric CXR images, categorised into pneumonia and normal classes, using the predefined patient-wise train, validation, and test splits to prevent data leakage and ensure fair evaluation. Consistent preprocessing and data augmentation techniques are applied to improve model generalisation. Each attention module is incorporated across all hierarchical stages of the network, enabling progressive refinement of feature representations from low-level visual patterns to high-level disease-specific characteristics. Model evaluation is conducted using accuracy, precision, recall, and F1-score, along with Grad-CAM visualisations to assess interpretability and highlight clinically relevant lung regions. An ablation study is conducted to analyse the contribution of attention placement across network stages and evaluate the impact of channel and spatial attention components. The experimental findings indicate that all attention-based models outperform the ResNet50 models, confirming the effectiveness of attention mechanisms in improving classification performance. ResNet50 achieved an accuracy of 91.03%, while ResNet50+SE, ResNet50+SA, and ResNet50+CBAM achieved 91.99%, 92.31%, and 92.95%, respectively. ResNet50+CBAM improved accuracy by 1.92 percentage points over ResNet50. Test results confirm this gain stems specifically from combined channel-spatial refinement across all stages, not either component alone. These findings offer the first controlled evidence that CBAM's advantage holds specifically for pediatric pneumonia detection.