University of Twente Student Theses
Seed Localisation for Quality Assurance of Seeding Patterns based on Image Analysis and Deep Learning
Visser, Okke (2024) Seed Localisation for Quality Assurance of Seeding Patterns based on Image Analysis and Deep Learning.
|
PDF
31MB |
| Abstract: | This project focuses on developing a prototype automated Quality Assurance (QA) system for analysing seed placement in gutters, using image processing and deep learning techniques to replace the current unquantifiable manual inspection. This type of practical application has no comparable research anywhere, and this project researched a unique solution for this problem combining multiple technologies. The system aims to provide real-time analysis on metrics like density, detect seed types on noisy backgrounds, and generate a comprehensive representative image for evaluation. A key challenge addressed in the project was the accurate detection of multiple seed types in varying substrates, focusing on optimising the YOLO-based object detection model for performance and precision. The system leverages image stitching to combine multiple images into a complete image representing the gutters while also integrating duplicate seed detection to ensure no duplicates on overlapping areas of the image. The prototype’s ability to generalise with new substrates was evaluated, showing promising results with minimal retraining required. The prototype successfully met the requirements, providing valuable data for seed analysis and showing potential for integration into the production environment. The results suggest that the system can be further enhanced with improved validation metrics, expanded seed datasets, and potentially more powerful hardware to support faster conveyor speeds. Future work will likely focus on refining the integration, enhancing model accuracy, and expanding the system’s ability to handle diverse substrates and seed types. |
| Item Type: | Essay (Master) |
| Clients: | Growy, Amsterdam, The Netherlands |
| Faculty: | EEMCS: Electrical Engineering, Mathematics and Computer Science |
| Subject: | 53 electrotechnology, 54 computer science |
| Programme: | Embedded Systems MSc (60331) |
| Link to this item: | https://purl.utwente.nl/essays/104756 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page