http://repository.ou.ac.lk/handle/123456789/4272 Fri, 09 Oct 2026 16:14:31 GMT 2026-10-09T16:14:31Z http://repository.ou.ac.lk/handle/123456789/4308 Plastic Subtype Recognition System Using a TinyML Based Lightweight CNN on ESP32 Smart Bin Ranasinghe, K.A.M.N.; Thenuwara, S.S.; Premachandra, H.W.H. Plastic waste segregation is important for optimizing the recycling process and minimizing the negative effects associated with improper disposal of plastic waste. Nonetheless, most of the smart waste management systems are designed mainly for general waste classification, fill level determination, or cloud computing, neglecting real time identification of different types of plastics at the disposal site. In this research, a plastic type recognition model based on edges is developed, embedded in a smart bin using the Light Weight Convolutional Neural Network and Tiny ML. The developed recognition model is able to classify plastic waste into five different types, namely PET, HDPE, LDPE, PP, and others/mixed. Hence, this result can be used to automatically sort the plastic waste to its correct compartment. The model was evaluated using a custom dataset created by combining images from multiple publicly available plastic waste image datasets. The lightweight MobileNetV2 architecture based CNN was trained with images of size 160 x 160 pixels (RGB format) and tested against the validation set. The trained model was then quantized to produce an INT8 TensorFlow Lite model to guarantee a smaller computation footprint for deploying the model in ESP32 hardware. While the trained Keras model gave an accuracy score of 92.08% on the validation data set, the quantized INT8 model produced an accuracy of 90.68% which translates to an approximate drop in accuracy of about 1.40 percent point. The implication is that with the help of lightweight deep learning models and INT8 quantization, one can still achieve an accurate classification of plastic subtypes. The conclusions drawn from this study is that TinyML based edge inference can be a practical way for plastic subtype classification in order to automate the process of smart waste segregation via classification without relying on cloud based classification continuously. Thu, 01 Jan 2026 00:00:00 GMT http://repository.ou.ac.lk/handle/123456789/4308 2026-01-01T00:00:00Z http://repository.ou.ac.lk/handle/123456789/4307 DESIGNING A STATABLE SAFARI WEAR COLLECTION FOR YOUNG FEMALE TOURISTS IN SRI LANKA Pathirathna, H.P.D.U.; Safana, H.K.H; Wijerathne, D.S. Sri Lanka is globally recognized for its wildlife, creating an increasing opportunity to engage young female tourists through fashion that respects both cultural and environmental aspects. As a result, a safari apparel line has been developed, drawing inspiration from the Sri Lankan leopard. This study examines the overlap of fashion and the safari tourism sector in Sri Lanka, concentrating on designing comfortable and modest travel garments for young women aged 25–30 from abroad who visit Sri Lanka as a safari location. The main goal of this project is to create a functional and visually appealing collection of safari clothing suitable for the spring and summer months in Sri Lanka while addressing the needs of sustainable applications for fabric material modification through setting biodegradable aurwedic fabric setting. A customer survey with 105 international female participants, including both past and potential visitors to Sri Lanka, revealed a demand for lack of a local safari apparel brand. These insights were instrumental in identifying the target demographic and guiding the design process. The theme was selected based on Sri Lanka's prominent position as a popular travel destination and current initiatives aimed at revitalizing its tourism sector. By connecting this collection to one of the nation's most significant natural attractions, the safari experience seeks to make a meaningful contribution to both the fashion industry and tourism. The Sri Lankan leopard, a native and endangered species, serves as the central influence, with its textures, rosette patterns, earthy hues, and agility shaping the design's visual and structural features. The collection incorporates breathable, skin-friendly materials like cotton twill to provide comfort in high UV conditions. Ayurvedic textiles are also included to reduce skin irritation and ensure sun protection. Moreover, methods such as screen printing, braiding with reeds, and separable garment components are utilized. Beyond just fashion, this project aims to foster environmentally friendly tourism. Thu, 01 Jan 2026 00:00:00 GMT http://repository.ou.ac.lk/handle/123456789/4307 2026-01-01T00:00:00Z http://repository.ou.ac.lk/handle/123456789/4306 Preventing Garment Waste at Source: Elastane Feeding Angle in Tights Knitting Kumasaru, D.S.D.; Priyakumari, M.A.D.S. Garments rejected during manufacture are an under-examined stream of textile waste. They carry the full environmental burden of fibre production, dyeing, knitting and finishing, but unlike post-consumer clothing they are discarded before delivering any service. Elastane-containing garments are particularly problematic, because elastane obstructs both mechanical and chemical recycling of blended textiles. This study asks whether elastane damage in knitted tights can be reduced at source by altering the geometry of the yarn path between creel and feeder. Quality-assurance records at a Sri Lankan hosiery manufacturer attributed 25% of elastane damage to the feeding system and, within that group, 44% to yarn path behaviour. Three feeding angles, 38.7°, 45.0° and 50.2°, were derived from a survey of 60 production machines and trialled with all other variables held constant. Screening over 100 pairs per angle gave defect rates of 1.0% at 45.0°, 1.5% at 38.7° and 5.0% at 50.2°. In seven days of continuous production, 45.0° recorded 6 defects in 720 pairs (0.83%) against 17 in 745 pairs (2.28%) at 38.7°, a statistically significant difference (χ2 = 4.08, df = 1, p = 0.043). Applied to the 351,042 pairs produced in 2025, of which 3,962 were lost to elastane damage, the trial rate corresponds to approximately 1,040 finished garments not discarded each year. The intervention requires repositioning a single yarn guide and no capital investment, indicating that process geometry is an accessible, under-used lever for waste prevention in apparel manufacturing. Thu, 01 Jan 2026 00:00:00 GMT http://repository.ou.ac.lk/handle/123456789/4306 2026-01-01T00:00:00Z http://repository.ou.ac.lk/handle/123456789/4305 Extraction and Characterization of Sri Lankan Habarala (Alocasia macrorrhizos (L.) G.Don) Fibre for Potential Composite Applications R.W, Tharusha Sithum Karunarathna; Balakrishnan, Subashini This study investigates the extraction and characterisation of natural fibre from Alocasia macrorrhizos (L.) G.Don, locally known as Habarala in Sri Lanka, with an emphasis on its potential as a lignocellulosic reinforcement for composite applications. Fibres were mechanically extracted from the stems using a decorticator, yielding an average fibre yield of 6.06%. The extracted fibres exhibited an average linear density of 11.57 Tex, fibre diameter of 31.19 μm, breaking force of 3.28 N, elongation at break of 1.74%, tenacity of 0.24 N/Tex, and moisture content of 14.83%. The mechanically extracted fibres were then subjected to three chemical treatment conditions (A1, B2, and C3) to evaluate changes in their physical, mechanical, and chemical characteristics. C3 produced the greatest reduction in linear density (26.53%) but also decreased breaking force and elongation at break, indicating deterioration of fibre structure. By comparison, B2 showed the best mechanical properties after treatment. FTIR analysis showed that the treatment removed hemicellulose and reduced lignin-related functional groups, while cellulose peaks became more prominent. The results suggest that mechanically extracted Habarala stem fibre is a promising lignocellulosic reinforcement for further evaluation in composite development. Thu, 01 Jan 2026 00:00:00 GMT http://repository.ou.ac.lk/handle/123456789/4305 2026-01-01T00:00:00Z